[Advances in Marine Biology] Aquatic Geomicrobiology Volume 48 || The Nitrogen Cycle

Download [Advances in Marine Biology] Aquatic Geomicrobiology Volume 48 || The Nitrogen Cycle

Post on 15-Dec-2016




0 download


  • 8.2. Nitrification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

    8.3. Nitrogen fixation and assimilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2661. INTRODUCTIONThe Nitrogen Cycle

    1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

    2. The Global Nitrogen Cycle and Human Perturbations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

    3. Biological Nitrogen Fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

    3.1. The nitrogenase enzyme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

    3.2. Ammonium assimilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214

    3.3. The oxygen problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215

    3.4. Phylogeny of nitrogen-fixing organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216

    3.5. Nitrogen fixation in aquatic environments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

    4. Microbial Ammonification and Nitrogen Assimilation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

    4.1. Ammonification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

    4.2. Deamination and ammonium incorporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

    4.3. Nitrogen mobilization and immobilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

    4.4. Anaerobic nitrogen mineralization and ammonium behavior in sediments . . . . . . 225

    4.5. New versus regenerated nitrogen in pelagic ecosystems. . . . . . . . . . . . . . . . . 230

    5. Nitrification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

    5.1. Biochemistry and thermodynamics of nitrification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

    5.2. Phylogeny of chemolithoautotrophic nitrifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

    5.3. Environmental factors aVecting nitrification rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

    6. Dissimilatory Nitrate Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246

    6.1. Biochemistry of denitrification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

    6.2. Biochemistry of NO3 ammonification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2516.3. Phylogeny and detection of denitrifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

    6.4. Environmental factors aVecting denitrification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

    7. Anammox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

    8. Isotope Fractionation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

    8.1. Denitrification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265Chapter 7Nitrogen is a key constituent of many important biomolecules, such as

    amino acids, nucleic acids, chlorophylls, amino sugars and their polymers,

    and it is essential to all living organisms. Nitrogen has the property of an

    ADVANCES IN MARINE BIOLOGY VOL 48 2005 Elsevier Inc.0-12-026147-2 All rights reserved

  • 206 CANFIELD ET AL.Figure 7.1 The microbial nitrogen cycle with an indication of the valence state ofthe various nitrogen-containing compounds involved.eight-electron diVerence between its most oxidized and reduced compounds(Figure 7.1). Thus, the redox cycling between nitrogen compounds forms

    the basis for numerous microbial metabolisms. Many of these microbial

    processes, in turn, control the availability of nitrogen in the environment

    and hence are significant in regulating the activities of primary producing

    eukaryotes, which require a ready source of nitrogen for growth.

    In a quick survey of the various microbial nitrogen-transforming process-

    es, we start with N fixation. In this process, prokaryotes transform atmo-

    spheric dinitrogen (N2) into ammonia (NH3), a biologically available form

    that can be incorporated into biomolecules (Sprent and Sprent, 1990). The

    diYculty in N fixation is breaking the strong triple bond holding the twoN atoms in N2 together. Most organic nitrogen is recycled into inorganic

    form by a process known as ammonification in which nitrogen-containing

    biomolecules are degraded by microorganisms or digested by animals. The

    released ammonium (NH4 ) can either be re-assimilated by microbes orplants and transformed into new biomolecules, or it can be oxidized by an

    assemblage of largely chemolithoautotrophic prokaryotes. The resulting

    oxidized nitrogen forms, nitrite (NO2 ) and nitrate (NO3 ), can either be

    assimilated by microorganisms and plants or be denitrified to N2 by a

    number of heterotrophic prokaryotes using oxidized nitrogen as electron

    acceptor and organic carbon, for example, as electron donor. The resulting

  • N2 is returned to the large atmospheric pool and thus, on short term, is lost

    for further biological transformations. Most bioavailable nitrogen is recycled

    several times between autotrophic and heterotrophic organisms, because the

    rates of nitrogen input to the biosphere by N fixation, and output by denitri-

    fication, are at least an order of magnitude slower than the internal cycling

    rates. This conventional view of the nitrogen cycle has recently been amended

    with the anammox process, in which ammonium oxidation is coupled to

    nitrite reduction, leading to the production of N2. This process may be of

    significance in the nitrogen cycle of aquatic enviroments.

    In this Chapter we explore how the interplay between microbial processes

    and the geochemical environment controls the cycling of nitrogen in aquatic

    ammonium substituted within potassium-rich minerals (Table 7.1) (Krohn

    THE NITROGEN CYCLE 207et al., 1988). Rock weathering liberates this nitrogen, which then becomes

    available to living organisms. Nitrogen in sediments and sedimentary rocks

    is the next largest pool. Here, too, nitrogen is mostly fixed as ammonium in

    secondary silicate minerals (Blackburn, 1983), and this nitrogen source also

    becomes available during weathering. Of comparable size is the reservoir of

    atmospheric N2, which accounts for 78% of the gas in the atmosphere. The

    Table 7.1 Major nitrogen reservoirs on earth

    Type of pool Location Pool size (g N)

    N2 gas Atmosphere 3.8 1021Living biomass (LPON) Aquatic and terrestrial 1.3 1016Dead organics (DPON + DON) Aquatic and terrestrial 9.0 1017Inorganic (NH4 , NO

    2 , NO

    3 ) Aquatic and terrestrial 2.4 1017

    Inorganic (fixed NH4 ) Sediments and sedimentary rock 4.0 1021Inorganic (NH4 within minerals) Igneous rock 1.4 1022

    From Delwiche, 1970; Blackburn, 1983; Madigan et al., 2002.The largest reservoir of nitrogen at the Earths surface, including the crust

    and the atmosphere, is in igneous rocks, where nitrogen is primarily found asenvironments. Thus, we look at how various microbial pathways promote

    the transformation of nitrogen compounds and how environmental factors

    regulate the operation and intensity of these pathways. In addition, we also

    explore how nitrogen isotopes are fractionated during microbial transfor-

    mation. First, however, we consider aspects of the global nitrogen cycle and

    the influence of human activities.


  • biologically available forms of fixed nitrogen mainly consist of dissolved

    NH4 , NO2 , and NO

    3 in aquatic and terrestrial environments, but this pool

    is small compared with the atmospheric reservoir (0.006%) and the reservoir

    comprising dead organic detritus (25%) (Vitousek et al., 1997). Livingbiomass is the smallest nitrogen reservoir, only about 1% of the size of the

    dead organic detrital pool.

    Pool sizes tend to be inversely related to their biological importance. The

    large igneous and sedimentary pools, for example, are not actively cycled by

    organisms, although rock weathering can contribute a locally significant

    source of nitrate to surface and ground waters (Holloway et al., 2001). The

    biological processes of N fixation and denitrification interact with

    the large atmospheric N2 pool (Figure 7.2), but even here, the turnover

    time of the atmospheric N2 pool is slow. The inorganic ions, NH4 , NO

    2 ,

    and NO3 , are distributed in aqueous solution throughout the ecosphere, andthey form small actively cycled reservoirs. Assimilation, mineralization,

    208 CANFIELD ET AL.Figure 7.2 The global nitrogen cycle with an indication of the most importanttransfer processes. Rates are given as 1012 g N yr1. NX indicates inorganic com-bined nitrogen. Based on data from Roswall (1983); Codispoti and Christensen(1985); Seitzinger and Giblin (1996); Vitousek et al. (1997); Seitzinger and Kroeze(1998). Other recent estimates of marine dentrification are as high as 450 Tg Ny1

    indicating a large possible imbalance of the marine N budget (Codispoti et al., 2001).

  • (Gruber and Sarmiento, 1997). Industrial nitrogen fixation into synthetic

    obvious that humans have heavily impacted the global nitrogen cycle. An-

    THE NITROGEN CYCLE 209other example of this impact comes from the cycle of nitric oxide (NO),

    which can be transformed in the atmosphere into nitric acid, a major

    component of acid rain (Vitousek et al., 1997). Fossil fuel burning emits

    more than 20 1012 g N yr1 as NO, while deforestation through burningcontributes another 10 1012 g N yr1 as NO. Furthermore, a substantialfraction of the total of 5 to 20 1012 g yr1 of NO nitrogen emitted fromsoils is human related. Overall, 80% or more of NO emissions worldwide are

    generated by human activities.

    To consider the ammonia cycle, nearly 70% of the global emissions to

    the atmosphere are human related. Ammonia volatilization from fertilized

    fields contributes an estimated 10 1012 g N yr1, release from domesticanimal wastes liberates about 32 1012 g N yr1, and forest burning con-tributes a further 5 1012 g N yr1. Synthetic nitrogen fertilizer input hasincreased fourfold over the last four decades (about 20 1012 g N yr1 in1960) and is expected to rise from a current level of 80 1012 g N yr1 to 1341012 g N yr1 in 2020 (Vitousek et al., 1997). Additional non-biologicalsources of fixed nitrogen to the Earths surface include volcanic outgasing

    and atmospheric fixation through ionizing radiation and electrical discharge.


    Only specialized prokaryotes contain the enzyme nitrogenase and the ability

    to fix N2 into a biologically useful combined form (Sprent and Sprent, 1990).

    These organisms, known as diazotrophs, are found in both prokaryote

    domains. They may be found both free-living and in symbiotic association

    with plants. Nitrogen-fixing prokaryotes utilize a wide variety of diVerentenergy metabolisms, including oxygenic photosynthesis, sulfate reduction,

    methanogenesis, anoxygenic photosynthesis, and chemolithoautotrophy

    (Table 7.2).fertilizer from the Haber-Bosch process produces roughly 80 1012 g yr1 offixed nitrogen, similar in magnitude to marine biological N fixation. It isand nitrification (internal cycles on land and in the ocean as shown in Figure

    7.2) are quantitatively the most important processes linking the inorganic

    reservoir with the likewise small and actively cycled reservoirs of living and

    dead organic nitrogen (JaVe, 2000).Biological nitrogen fixation on land amounts to about 180 1012 g yr1,

    which includes about 40 1012 g yr1 from agriculturally managed legumecrops (Figure 7.2). This rate exceeds rates of biological nitrogen fixation in

    the marine environment, with an estimate of about 100 1012 g yr1

  • Table 7.2 Selected list of free-living microbial genera, which contain N-fixingspecies or strains. The list is separated into diVerent metabolic types of microorgan-

    210 CANFIELD ET AL.isms with indication of their preferred habitat

    Metabolism Genus or type Environment

    Aerobes Azotobacter Sediment/waterFacultative anaerobes

    (not fixing when aerobic)Klebsiella Sediment/waterPaenibacillus Microbial mat/rhizosphereEnterobacter Sediment/animal gutEscherichia Animal gut

    Microaerophiles(when fixing N2)

    Xanthobacter Microbial mat/sedimentThiobacillus Microbial mat/sedimentAzospirillum RhizosphereAquaspirillum Water

    Anaerobes Clostridium SedimentDesulfovibrio SedimentMethanosarcina SedimentCombined, a major nutrient for plant primary production, is generally

    found in short supply in marine areas, where significant loss of occurs

    through denitrification (see Section 6). Nitrogen is also lost through burial

    as organic N and as adsorbed and mineral-bound ammonium in sediments

    (Figure 7.3). Nitrogen fixation is therefore needed to replace lost nitrogen.

    The overall process of N fixation is exergonic at standard conditions (Equa-

    tion 7.1), but the great stability of the NN bond in N2 makes it extremelyunreactive at room temperature.

    3H2 N2 ! 2NH3; DG 0 33:4 kJ mol1 7:1Indeed, in the industrial Haber-Bosch process this reaction will only occur

    when operated at temperatures between 300 and 4008C and at pressures

    Methanococcus SedimentPhototrophs Chromatium Microbial mat/sediment

    Chlorobium Microbial mat/sedimentThiopedia Microbial mat/sedimentRhodospirillum Microbial mat/sedimentRhodopseudomonas SedimentOscillatoria Microbial mat/waterNodularia WaterAnabaena WaterNostoc Microbial matCalothrix Microbial matGloeotheca Microbial mat

    Modified from Capone (1988) and Postgate (1998).

  • THE NITROGEN CYCLE 211between 35 and 100 MPa. Prokaryotes outperform the industrial pro-

    cess with the nitrogenase enzyme complex. Thus, nitrogenase reduces the

    triple-bonded NN molecule to NH4 at normal environmental tempera-tures and pressures, but it is not completely N2 specific, as other triple-

    bonded molecules such as acetylene (HCCH) and hydrogen cyanide(HCN) are also reduced.

    Figure 7.3 Microbial nitrogen cycling in aquatic environments showing themajor transformations within (uppercase letters) and between (lowercase letters)anoxic sediment, oxic sediment, the water column, and the atmosphere. A, nitrogenfixation; B, NOx assimilation; C, ammonification; D, NH

    4 assimilation; E, NH


    oxidation; F, NO2 oxidation; G, NO3 ammonification; H, denitrification; I, ana-

    mmox; a, burial; b, downward diVusion; c, upward diVusion; d, NH4 adsorption;e, NH4 desorption.

  • Measuring N fixation

    toxicological, and enzymological factors may result in deviations from this

    ideal stoichiometry. For N-fixation rates measured in aquatic sediments, the

    212 CANFIELD ET AL.ratio of C2H2 reduction to N2 reduction may range from 10:1 to 100:1

    (Seitzinger and Garber, 1987). Ideally, the C2H2 reduction assay should be

    calibrated against 15N2 uptake measurements before it can faithfully provide

    quantitative measurements of N fixation (Seitzinger and Garber, 1987).

    3.1. The nitrogenase enzyme

    Nitrogenase is large (up to 300 kDa) compared with many other enzymes,

    and relatively slowly reacting, with one enzyme taking 1.25 s to form two

    NH4 molecules (Equation 7.2) (Postgate, 1998). The formation and mainte-nance of the nitrogenase enzyme complex require a major investment of

    protein (up to 30% of total cell protein; Haaker and Klugkist, 1987), energy

    (ATP), and trace metals (Mo and Fe). The trace metals incorporate into

    the active core of two diVerent components associated with nitrogenase, aMo-Fe protein known as dinitrogenase and an Fe protein known as dinitro-

    genase reductase. Additionally, some organisms have Mo-free dinitrogenases

    with cofactors containing V and Fe, or Fe only, which are structurally and

    functionally similar to the Mo-Fe dinitrogenase but appear to be less eYcient(Eady, 1996). The Fe-only proteins may have been of particular significance

    through much of the Proterozoic Eon of Earth history (2.5 to 0.54 billionThe development of the acetylene (C2H2) reduction assay (Stewart et al.,

    1967; Hardy et al., 1968) revolutionized the study of N fixation by providing a

    sensitive and simple procedure for determining nitrogenase activity. The pro-

    cedure relies on the low specificity of nitrogenase for its natural substrate (N2)

    and its capacity to reduce other triply bonded small molecules. Acetylene is

    reduced in preference to N2 by nitrogenase, rapidly forming ethylene (C2H4):

    HC CH 2H 2e ! H2C CH2The C2H4 generated in the C2H2 reduction assay is quantified by gas chroma-


    There are, however, some potential concerns with this assay. Most impor-

    tantly, the concentrations of C2H2 required in the assay may be inhibitory to a

    broad range of both diazotrophs and non-diazotrophs (Capone, 1988). Anoth-

    er concern relates to stoichiometry of the reaction. In principle, the two

    electrons transferred during the reduction of C2H2 to C2H4 are three times

    less than the six electrons required to reduce N2 to NH4 . Thus, three moles of

    C2H2 reduced should be equivalent to one mole of N2 fixed. Physiological,

  • THE NITROGEN CYCLE 213years ago), when seawater Mo concentrations were probably quite low

    (Anbar and Knoll, 2002).

    The Fe protein consists of two identical subunits coded for by the nifH

    gene and contains approximately four Fe atoms and four labile S atoms. The

    Mo-Fe protein consists of four subunits, which are pairs of two diVerent

    Figure 7.4 Schematic representation of the nitrogenase system, which catalyzesthe reduction of molecular nitrogen to ammonia. The enzyme system containsdinitrogenase reductase (Fe protein) and dinitrogenase (Mo-Fe protein).types: the a subunit coded for by the nifD gene and the subunit coded forby the nifK gene. The Mo-Fe protein complex contains two Mo, 2135 Fe

    and 1824 labile S atoms per molecule (Postgate, 1998).

    The reduction of N2 to NH4 proceeds as follows (Figure 7.4): a low redox

    potential molecule such as ferredoxin (at least 430 mV) donates an electronto the Fe protein, which enables it to react with MgATP. Meanwhile, the N2molecule to be reduced combines with the Mo-containing part of the Mo-Fe

    protein. The two proteins now join to form the active enzyme complex.

    Electrons flow singly from the Fe protein to the Mo-Fe protein. Thus, the

    two proteins have to meet and separate eight times to accommodate the six

    electrons transferred to reduce one N2 molecule to two NH4 molecules and

    the two electrons needed to reduce two protons to H2 (see Figure 7.4). The

    formation of H2 during the reduction of N2 is an inherent property of the

    enzyme. Overall, the N fixation reaction, including its cost in ATP, may be

    written as the following:

    N2 9H 8e 16ATP ! 2NH4 H2 16ADP 7:2

  • 3.2. Ammonium assimilation

    The ammonium formed during N fixation is incorporated into cellular

    material by one of two pathways (Gottschalk, 1986). Most nitrogen-fixing

    organisms produce glutamate as their initial product of NH4 assimilation,and in one pathway glutamate is formed from the reductive amination of

    a-oxoglutarate by the enzyme glutamate dehydrogenase (GDH):

    NADPHNH4 a-oxoglutarate ! NADP glutamateH2O7:3

    GDH has only a moderate aYnity for NH4 , and since NH4 generally represses

    nitrogenase activity, a rapid assimilation is imperative. Thus, an alternative

    two-step high-aYnity assimilation pathway that maintains a low NH4 concen-tration is sometimes used (Figure 7.5). Initially, NH4 is added to glutamate byan ATP-requiring step with the enzyme glutamine synthetase (GS):

    214 CANFIELD ET AL.glutamateNH4 ATP ! glutamineADP 7:4Glutamine is transformed back into glutamate by a second reaction, catalyzed

    by the enzyme glutamate synthase or GOGAT (glutamine-a-oxoglutarate-amino-transferase):

    glutamine a-oxoglutarateNADPH ! 2 glutamateNADP 7:5The cost of theGS-GOGATpathway is oneATPper glutamate formed, but the

    benefit to the organism is rapid assimilation of NH4 from low concentrationsin the environment, preventing suppression of nitrogenase activity.

    Figure 7.5 Assimilation of ammonia by the glutamine synthetase/glutamatesynthase (GS/GOGAT) enzyme complex. Modified from Gottschalk (1986).

  • THE NITROGEN CYCLE 2153.3. The oxygen problem

    The biochemical properties of nitrogenase present the organism with a

    number of physiological problems. Most important is the O2 sensitivity of

    the two main proteins. Thus, both the Mo-Fe protein and the Fe protein are

    irreversibly inactivated by O2. All diazotrophs are therefore obliged to

    protect nitrogenase from exposure to O2. Numerous strategies have been

    adopted to avoid O2, including (1) life without O2, (2) high rates of O2respiration, (3) conformational protection, and (4) heterocyst utilization

    (Postgate, 1998; Madigan et al., 2003). Thus, obligate anaerobes such as

    Clostridium pasteurianum and Desulfovibrio desulfuricans can fix nitrogen

    whenever their normal metabolism is active. Facultative anaerobes such as

    Klebsiella spp. and Enterobacter spp. are only capable of diazotrophic

    growth under anoxic conditions and must rely on other nitrogen sources in

    the presence of O2.

    Aerobic (e.g., Azotobacter spp.) and microaerophilic (e.g., Azospirillum

    spp.) diazotrophs can fix nitrogen only when their respiration decreases the

    O2 concentration near the cell to low enough levels to protect the functioning

    enzyme. Therefore, a characteristic feature of aerobic diazotrophs is excep-

    tionally high O2 respiration rates. As an example, Azotobacter spp. have cell-

    specific O2 uptake rates several times higher than that needed for normal

    catabolism (Poole and Hill, 1997). These high respiration rates, however,

    reduce O2 levels to the point where nitrogenase can fix N2 (see also Chapter 6,

    Section 5.2). Some Azotobacter species are also able to maintain undamaged

    nitrogenase under oxic conditions by conformational protection. In this

    case, nitrogenase proteins undergo a reversible conformational change to a

    state that is unaVected by O2. In this way, the enzyme becomes inoperativebut remains undamaged, and can readily resume nitrogen fixation as soon as

    O2 disappears.

    Some phototrophic cyanobacteria, which live in oxic environments and

    generate O2 in the light, have solved the problem of nitrogenase inactivation

    by separating photosynthesis and nitrogen fixation in space and time. Cya-

    nobacteria such as Anabaena cylindrica have specialized thick-walled cells

    called heterocysts at regular intervals along their filaments. Nitrogenase

    activity is normally restricted to the heterocyst, which lacks an oxygen-

    evolving apparatus. The heterocyst walls allow N2 to diVuse into the cellbut limit O2 diVusion to the point where any oxygen entering the heterocystcan be scavenged by respiration (Walsby, 1985). Some non-heterocystous

    cyanobacteria conduct nitrogen fixation at night, when O2 generation can-

    not occur. The unicellular cyanobacteria Gloeotheca spp. can fix nitrogen in

    the light and dark, but when grown in alternating lightdark cycles it fixes 20

    times more nitrogen in the dark than in light (Postgate, 1998). In this type of

  • Nitrogen fixation is found among members of the Bacteria and the

    216 CANFIELD ET AL.Euryarchaeota kingdom of the Archaea (Young, 1992). Within these groups,

    nitrogenase activity has been reported, or inferred, in more than 100

    genera distributed over many of the major phylogenetic divisions. Such a

    wide distribution suggests that the process has a very ancient origin (see also -

    Berman-Frank et al., 2003).Phylogenieshavebeen constructed from16S rRNA

    and from several of the diVerent nif genes associated with nitrogenase. In thissection we discuss phylogenies from the nifH gene, a principal gene in the

    N fixation pathway, and compare these with 16S rRNA phylogenies.

    The nifH phylogenetic tree organizes into several clusters (Figure 7.6),

    including the cyanobacteria, the actinomycetes, the a-proteobacteria, andthe and -proteobacteria. All of these are aerobes. There are also twoclusters of anaerobes, one including members of the Bacteria and the other

    including members of the Archaea. In addition, the aerobic gram-positive

    genus Paenibacillus clusters with the cyanobacteria and is well separated

    from the anaerobic members of gram positives represented by the genus

    Clostridium. In general, the nifH phylogeny is concordant with the 16S

    rRNA phylogeny. This supports the thesis that, for the most part, nitroge-

    nase genes have descended vertically from an ancient progenitor (Chien and

    Zinder, 1994, 1996). Therefore nifH evolution, as well as the evolution oforganism, and in other non-heterocystous cyanobacteria, O2 is removed to

    low concentrations by rapid respiration in cell colonies.

    Members of the cyanobacterial genus Trichodesmium are the most important

    nitrogen fixers in the ocean. Trichodesmium spp. are filamentous and non-

    heterocystous with individual cells organized along trichomes, which frequent-

    ly bundle into large aggregates. Nitrogenase is found only in a fraction of the

    cells, yet the oxygen-evolving apparatus is found in all cells, and at any given

    moment during the day, when N fixation occurs, a significant fraction of the

    nitrogenase is inhibited by oxygen (Berman-Frank et al., 2003). To accomplish

    nitrogen fixation, Trichodesmium spp. temporally separates nitrogen fixation

    and oxygen production (Berman-Frank et al., 2001b). Nitrogen fixation is

    concentrated at midday, when there is a lull in oxygen production. Also at

    midday there is an increase in light-dependent oxygen consumption by the

    Mehler reaction (see Chapter 6). The combination of low oxygen production

    and high rates of oxygen consumption around noon reduce the oxygen con-

    centrations in the cell and allow for nitrogenase activity. The rather contradic-

    tory occurrence of daytime N fixation by Trichodesmium spp. occurs because of

    the careful temporal regulation of N fixation and oxygen production and the

    high rates of oxygen consumption by the light-dependent Mehler reaction.

    3.4. Phylogeny of nitrogen-fixing organisms

  • THE NITROGEN CYCLE 217other nif genes (Fani et al., 2000), is due mostly to the constrained evolution

    of an ancient gene rather than lateral gene transfer. However, the placement

    of Paenibacillus in the nifH phylogeny is inconsistent with its gram-positive

    location in the 16S rRNA phylogeny and points to the possibility of lateral

    gene transfer of this gene.

    There are some other unusual aspects of nifH gene phylogeny. Some

    Bacteria, such as Rhodobacter capsulatus, Desulfobacter curvatus, and

    Clostridium pasteurianum, contain multiple copies of the nifH gene. In each

    of these cases, one nifH copy clusters with its natural neighbors based on 16S

    rRNA, and the other clusters elsewhere. For example, one of the nifH genes

    for both D. curvatus and C. pasteurianum clusters with the methanogens

    (Braun et al., 1999). Multiple nifH gene copies may have arisen during an

    ancient gene duplication event, as is apparently the case for other nif genes

    (Fani et al., 2000). If so, it is unclear why multiple nifH genes from the same

    Figure 7.6 Phylogeny constructed for amino acid sequences from a fragmentof the nifH gene, analyzed by the neighbor-joining method. Scale bar indicates2.9% sequence divergence. Modified from Achouak et al. (1999).

  • organism followed such diVerent phylogenic trajectories. Also unclear iswhether the multiple nifH genes actually accomplish the same function.

    3.5. Nitrogen fixation in aquatic environments

    Cyanobacteria are responsible for most planktonic nitrogen fixation in open

    marine and lake waters, but rates are high only when cyanobacteria domi-

    nate the planktonic biomass, which occurs frequently in eutrophic lakes and

    estuaries (Table 7.3). Nitrogen-fixing cyanobacteria need not be free living.

    In wetlands, for example, cyanobacteria such as Anabaena spp. live in a

    symbiotic relationship with the floating water fern Azolla spp., and in this

    relationship nitrogen fixation rates are greatly enhanced. In oligotrophic and

    mesotrophic lakes as well as the open ocean, rates of pelagic nitrogen

    fixation tend to be very low. Thus, the role of nitrogen fixation in supplying

    N to primary producers in open water ecosystems is quite variable and

    usually low (generally

  • and estuaries, but may be high in particularly organic-rich estuarine sediments

    fixation in oligotrophic lakes and shallow coastal systems is often dominated

    THE NITROGEN CYCLE 219by cyanobacteria. Benthic nitrogen fixers are generally unimportant for the

    nitrogen budget of mesotrophic and eutrophic lakes and estuaries, but may be

    an important nitrogen source in nutrient-poor tropical marine lagoons and

    oligotrophic lakes. Nitrogen fixation in wetlands (e.g., salt marshes and man-

    grove forests) and seagrass beds (e.g., Zostera sp. and Thalassia sp.) may be

    several-fold higher than in comparable unvegetated sediments (McGlathery

    et al., 1998; Nielsen et al., 2001). This is due to stimulated nitrogen fixation in

    the rhizosphere fueled by labile root exudates derived from plant photosynthe-

    sis. However, the contribution of newly fixed nitrogen in these very productive

    vegetated ecosystems is, in most cases, small relative to the total nitrogen input.

    Nitrogen fixation, therefore, appears to be important in making up deficits in

    nitrogen availability, relative to phosphorus, in many lakes and possibly the

    open ocean,whereasmany estuaries and coastal seas are nitrogen limited due to

    the relatively low rates of nitrogen fixation found in these productive systems.


    4.1. Ammonification

    Nitrogen in living and dead organic matter occurs predominantly in the

    reduced amino form, principally in proteins and nucleotides (e.g., DNA,

    RNA, and ATP). On average, prokaryote cells contain about 55% proteins

    and 23% polynucleotides, on a dry weight basis (Madigan et al., 2003). When

    these compounds are hydrolyzed and catabolized by heterotrophic organisms,

    nitrogen is ultimately liberated in the form of NH4 (ammonification, alsoknown as nitrogen mineralization), which can be assimilated and incorporated(Table 7.3). In mesotrophic and eutrophic lakes as well as estuarine sediments,

    most benthic nitrogen fixation is mediated by heterotrophic and chemolithoau-

    totrophic prokaryotes in the absence of light. However, benthic nitrogenalleviate nitrogen limitation. By contrast, current N budgets for the marine

    realm suggest that N-fixing organisms are not active enough to balance N

    losses by denitrification, and hence the oceans may be losing fixed nitrogen

    (Figure 7.2). However, past estimates of oceanic nitrogen fixation rates may

    have been low by an order of magnitude or more (Gruber and Sarmiento,

    1997; Karl et al., 1997). At the same time, global estimates of denitrification

    are rising, so the state of balance of the present-day marine nitrogen budget is

    still in doubt (Codispoti et al., 2001).

    Rates of nitrogen fixation are low to moderate in the sediments of most lakes

  • teins and polynucleotides yields oligopeptides, amino acids, oligonucleotides,

    220 CANFIELD ET AL.and nucleotides, all having C:N ratios below 6. Subsequently, intracellu-

    lar fermentative and respiratory processes deaminate these small dissolved

    molecules, resulting in the release of NH4 .Ammonification occurs in both oxic and anoxic aquatic environments

    (Herbert, 1999). Oxic environments include most lakes, swamps, rivers,

    and marine water bodies as well as their underlying near-surface sediments.

    Anoxic environments include euxinic and stratified water bodies and

    subsurface sediments. The partitioning between aerobic and anaerobic am-

    monification in aquatic environments depends, to a large measure, on

    the depth of the oxic water column. This controls the time elapsed before

    organic particles introduced into the photic zone reach anoxic environ-

    ments, either by sinking into anoxic water masses or by burial into anoxic

    sediment. In shallow coastal environments overlying intensively bioturbated

    sediments, about 50% of the ammonification occurs in the oxic water

    column, and the remainder occurs by both aerobic and anaerobic proces-

    ses within the sediment. The balance between aerobic and anaerobic ammo-

    nification varies, but 50% or more anaerobic ammonification is found in

    some cases. In the deep sea, more than 99% of the organic nitrogen in sinking

    particles usually is mineralized in the oxic water column (Suess, 1980).

    Ammonification always occurs in conjunction with heterotrophic carbon

    mineralization. Because nitrogen-containing polymers are generally degraded

    more easily than carbon-containing structural cell components such as cellulose

    and lignin, nitrogen is mineralized preferentially to carbon. As a consequence,

    the average C:N ratio of decomposing organic matter increases gradually

    during microbial decomposition (Blackburn and Henriksen, 1983). Such pref-

    erential removal of nitrogen during the early aerobic stages of decomposition

    provides less degradable and nitrogen-poor organic matter for later anaerobic

    decomposition (Figure 7.7). The increase in C:N ratio may occur over short

    depth scales in sediments, compared to the water column, because sediment

    accretion rates are slow compared to the sinking rates of particles in water.

    4.2. Deamination and ammonium incorporation

    The first step in the ammonification of amino acids is enzymatic deami-

    nation, in which amino groups are released and transformed into ammonium.

    This is accomplished by a diverse range of aquatic microorganisms, includingback into biomolecules by a variety of aerobic and anaerobic organisms.

    Alternatively, the ammonium can be oxidized by nitrifying microorganisms

    under oxic conditions (Figure 7.3) (JaVe, 2000).Aswith carbon (see Chapter 5),the rate-limiting step in nitrogen mineralization is the extracellular hydrolysis

    of macromolecules into smaller units. The initial microbial hydrolysis of pro-

  • Figure 7.7 Hypothetical change in C:N ratio of organic matter with progressiveorganic matter mineralization (expressed here as depth) in aquatic environments.Organic matter with a C:N molar ratio of about 6.6 (Redfield, 1958) is produced byphytoplankton in the photic zone of the water column. Organic matter sinking belowthe photic zone becomes nitrogen deplete as nitrogen-rich compounds such asproteins and nucleotides are degraded preferentially to structural carbon-rich com-pounds. The rate of nitrogen stripping decreases gradually with depth, in concertwith the decreasing content of labile and nitrogen-rich compounds. Particles reachingthe sediment surface typically have a C:N ratio of 810 (Richardson, 1996). Thepreferential nitrogen mineralization continues within the sediment, but at a muchlower rate. The C:N ratio of organic matter increases to a level of up to 15 within theupper tens of centimeters of sediment.


  • members of the genera Pseudomonas, Vibrio, Proteus, Serratia, Bacillus, and

    222 CANFIELD ET AL.Clostridium, as well as many actinomycetes and fungi (Herbert, 1999).

    Amino acid deamination primarily occurs as an oxidative process catalyzed

    by NAD-linked dehydrogenases (Gottschalk, 1986):


    Alternatively, oxidative deamination in respiring organisms can occur as a

    reaction with oxygen. Deamination of amino acids can also occur as a non-

    oxidative process catalyzed by dehydratases and is associated with the

    removal of water.

    The hydrolysis and deamination of nucleotides are more complex due to

    the heterocyclic structure of the molecules. The degradation of purines and

    pyrimidines is not well studied, but the process is assumed to occur via either

    an oxidative pathway, where both urea and NH4 are released, or via areductive pathway, where only NH4 (and probably amino acids) is producedas shown below for cytosine (Vogels and van der Drift, 1976; Gottschalk,

    1986; Therkildsen et al., 1996):


    Urea formed by the oxidative pathway is further mineralized to CO2 and

    NH4 by the action of the enzyme urease according to the following:


    The NH4 ions released during mineralization can be assimilated into aminoacids, nucleotides, and other nitrogen-containing biomolecules by numerous

    plants and microorganisms. The biochemistry of NH4 incorporation intoorganic matter is similar to what was previously described for nitrogen

    fixers. In brief, when the concentration of NH4 in the environment ishigh it is assimilated by reductive amination of a-oxoglutarate, formingglutamate by the action of the enzyme glutamate dehydrogenase (GDH)

    (Gottschalk, 1986). It is clear that GDH, with a Km for NH4 of 0.1 M,

    cannot be involved in the assimilation of NH4 in most aquatic environ-ments, where the concentration of NH4 usually is low (1mM). Undersuch conditions, a combination of glutamine synthetase and glutamate

    synthase is responsible for glutamate formation from a-oxoglutarate and

  • THE NITROGEN CYCLE 2234.3. Nitrogen mobilization and immobilization

    The free NH4 pool is generally small, but dynamic, in aquatic environmentswith high production and consumption rates. Microbial mineralization

    of organic nitrogen occurs in conjunction with microbial incorporation

    (assimilation) of soluble inorganic nitrogen. The balance between theseMeasuring ammonification and nitrogen incorporation

    Temporal changes in the NH4 pool reflect net ammonification or net nitro-gen incorporation. However, when more precision is required, or when gross

    rates of NH4 turnover are desired,15NH4 methods can be used (Blackburn,

    1979a). Thus, by assaying the change in NH4 pool sizes and15N dilution

    through time after adding 15NH4 to the NH4 pool, gross rates of ammonifica-

    tion (Nd) and nitrogen incorporation (Ni) can be quantified with model calcu-

    lations. Dilution of 15NH4 during incubation indicates that the label has beenincorporated into new biomass or has been diluted by ammonification. Net

    ammonification (No Nd Ni) can be calculated from the temporal change inNH4 concentration (Ct):

    Ct Not C0Based on the initial NH4 concentration (C0) and labelling percentage (%


    and the NH4 concentration (Ct) and labelling (%15Nt) at time t, the following

    relationship can be established:

    ln%15Nt ln%15N0 Nd=No ln Ct=C0

    A plot of ln(Ct/C0) against ln(%15Nt) provides a slope of Nd/No, from which

    the rate, Nd, is easily calculated by the use of the measured No. Ni can then be

    calculated from (Ni NdNo; see above). When the method is applied to oxicenvironments the incorporation term, Ni, also includes

    15NH4 , which is lost byoxidation.

    One major problem arises when this method is used in sediments. A variable

    amount of 15NH4 disappears immediately into the sediment matrix, andcannot be extracted with KCl or other salts, which are commonly used to

    liberate adsorbed NH4 from sediment particles.NH4 . In this pathway, glutamate synthase (GOGAT) catalyzes the reduc-tive transfer of the amino group of glutamine to a-oxoglutarate. Gluta-mate is always the initial amino acid formed in the assimilation process.

    From this compound, the amino group can be transferred through trans-

    amination to other a-oxoacids, thus forming their corresponding aminoacids.

  • two processes (mineralization and incorporation) is determined by a variety

    224 CANFIELD ET AL.of factors, including the ease of hydrolysis, the chemical composition of the

    organic compounds, the growth rates of the organisms, and environmental

    factors such as redox conditions. Fresh protein- and polynucleotide-rich

    organic materials derived from phytoplankton (bulk C:N 7) and animaltissues (C:N 5) are readily hydrolyzed and mineralized to CO2 and NH4 ,providing an excess of dissolved inorganic nitrogen. A large proportion of

    this liberated NH4 is mobilized into the environment. By contrast, duringthe initial microbial hydrolysis and metabolism of structural carbohydrates

    such as cellulose-rich tissues from vascular plants poor in nitrogen (C:N >20) there is a demand for incorporation of dissolved inorganic nitrogen

    into cell biomass. Little of the nitrogen may be mobilized. In later stages

    of decomposition and population growth, when most of the available

    carbon is oxidized and the microbial population declines, the mineralization

    of nitrogen-rich bacterial biomass can provide excess nitrogen to the

    environment (Figure 7.8).

    The intimate association between organic carbon and nitrogen and the

    balance between immobilization and mobilization of nitrogen during micro-

    bial mineralization can be visualized with a C:N ratio model (modified

    from Blackburn, 1979b). This non-steady-state model considers the relation-

    ship between the C:N ratio of the organic substrate used by heterotro-

    phic microbial populations (C:NS), the C:N ratio of the growing microbial

    cells (C:NB), and the eYciency of carbon incorporation (EC) into newcells (Figure 7.9). Organic carbon is degraded at a rate of Cd Co Ci,where Co is the rate of oxidation to CO2 and Ci is the rate of incorpora-

    tion into microbial cells. Similarly, organic nitrogen is degraded at a rate of

    Nd No Ni, where No is the rate of net mineralization and Ni is therate of incorporation. Since the nitrogen transformations are diYcultto measure, they can be deduced from carbon transformations as follows:

    EC Ci=Cd;Nd Cd=C:NS and Ni ECCi=C:NB. When all nitrogen as-similated by microbes is in the form of NH4 , then the C:N ratio of substrates(C:NS) and the mobilization/immobilization of nitrogen, MIN Ni/Nd, arerelated according to the following equation:

    MIN ECC:NS=C:NB 7:9

    Mobilization occurs when MIN < 1 and immobilization occurs whenMIN > 1. If it is assumed that EC is 0.5 for aerobes and 0.2 for anaerobes(Pedersen et al., 1999), and that C:NB is 5 (Goldman et al., 1987), then

    nitrogen mobilization will occur only when C:NS is below 10 under

    oxic conditions and below 25 under anoxic conditions (Figure 7.10).

    These predictions are in close agreement with observations from laboratory

  • THE NITROGEN CYCLE 225experiments on cultured aerobic marine prokaryotes (Goldman et al., 1987;

    Tezuka, 1990) and on anaerobic sediment microbes (Blackburn, 1979b).

    4.4. Anaerobic nitrogen mineralization and ammonium behaviorin sediments

    Nitrogen mineralization in sediments is divided between the upper oxic layer

    and the deeper anoxic layer, which in coastal and shelf sediments can be

    of particular significance due to the generally shallow depths of oxygen

    Figure 7.8 Hypothetical changes in carbon and nitrogen pools with time indegrading plant materials. (A and C) Decomposition of diatoms (initial C:N ratioof 7); (B and D) decomposition of vascular plant material (initial C:N ratio of 50). (Aand B) Temporal patterns of carbon (broken lines) and nitrogen (full lines) in thedegrading materials given as percent of the initial content; (C and D) percent of theinitial nitrogen content in the degrading materials that have been mobilized orimmobilized per day.

  • Figure 7.10 The relationship between C:N ratio of substrates (C:NS) and MIN(see text for explanation). Two examples are given: when carbon incorporationeYciency (EC) is 0.5 (aerobic prokaryotes) and 0.2 (anaerobic prokaryotes). TheC:N ratio of prokaryotes (C:NB in text) is fixed at 5. The intercept between MIN 1(dotted line), and regression lines represent the conditions at which no net exchangeof NH4 occurs. The symbols represent measured values from laboratory experimentson cultured marine aerobes (filled circle; Goldman et al., 1987) and on sedimentanaerobes (filled triangle; Blackburn, 1979b).

    Figure 7.9 Schematic presentation of carbon and nitrogen mineralization with anindication of the terms used in the C:N ratio model of Blackburn (1979b). See text formore details.


  • penetration in these environments (see Chapter 6). Anoxic pathways of

    nitrogen mineralization depend little on the terminal carbon mineralization

    pathway. Thus, while we think of sulfate reduction and methanogenesis as

    important pathways of carbon mineralization in anoxic marine and freshwa-

    ter sediments, nitrogen mineralization is largely uncoupled from the terminal

    carbon oxidation process. In anoxic sediments, nitrogen mineralization

    occurs during the initial hydrolysis and/or fermentation step of anaerobic

    decomposition (Kristensen and Hansen, 1995). Large biomolecules are fer-

    mented to small nitrogen-poor organic molecules (fatty acids such as acetate),

    while amino acids and nucleotides are deaminated to NH4 and fatty acids(Figure 7.11). The overall coupling between carbon oxidation and nitrogen

    mineralization (i.e. CO2 and NH4 production), as typically observed in

    THE NITROGEN CYCLE 227Figure 7.11 Decomposition products from anoxic marine sediment using a plugflow-through reactor system (as developed by Roychoudhury et al., 1998). Theconcentration scale (D concentration) shows the diVerence between water exitingand entering the sediment and represents the net carbon and nitrogen transforma-tions by the anaerobic community. When unamended, the sediment produced0.4mM CO2, 0.06 mM NH

    4 , and no DOC. Alanine and acetate added at a concen-

    tration of 1 mM C in the inflowing water was mineralized almost completely by themicrobial community with only traces of DOC. The mineralization products reflectthe stoichiometry of the substrates (1 mM CO2 and for alanine 0.33 mM NH

    4 ).

    Sulfate reduction can be inhibited by addition of 10mM molybdate. When molyb-date (+Mo) was added together with alanine, almost all carbon was recovered inorganic form as DOC with only traces of CO2, whereas all nitrogen was mineralizedcompletely to NH4 . Data from K. S. Hansen (unpublished).

  • sediment porewaters or incubation experiments (e.g., Canfield et al.,

    1993a,b), is therefore rather indirect and a result of the rapid turnover of

    the fermentation products.

    Detritus is rapidly decomposed in shallow sediment depths, and conse-

    quently more than 80% of the total nitrogen mineralization (Nd) generally

    occurs within the upper 0.1 m of the sediment (Figure 7.12). Whereas nitro-

    gen incorporation rates typically decrease with depth, the percentage of the

    total mineralized nitrogen (Nd) incorporated into biomass (Ni) may increase,

    as shown in Figure 7.12. In this case, incorporation ranges from 10% at

    02 cm to 40% at 1214 cm. This increase with depth is due to decreasing

    nitrogen content (or increasing C:N ratio) of the organic detritus and lower

    degradability of the detritus. The microorganisms assimilate and incorporate

    more of the mineralized nitrogen to maintain growth in the deeper layers if

    carbon incorporation eYciency (EC) remains constant.The concentration of NH4 in anoxic sediments is always higher than in

    the overlying water when there is net production. The produced NH4diVuses, or is advected by animals or bottom currents, to the sedimentsurface. Here it can be oxidized by nitrifying bacteria (see Section 7.5),

    assimilated by benthic microalgae (when there is light) or released to the

    overlying water. Owing to both nitrification and assimilation, NH4 fluxes to

    228 CANFIELD ET AL.the overlying water are generally lower than the total net production rates

    (No) in the sediment (Table 7.4).

    Figure 7.12 Depth dependence of NH4 production in a coastal sediment from theLimfjord, Denmark. (A) Indication of how much of the mineralized N is liberated tothe surroundings (No) and how much is incorporated into organisms (Ni). (B)Concentrations of NH4 in the porewater (Pw) and adsorbed to particles (adsorbed).Modified from Blackburn (1979a).

  • THE NITROGEN CYCLE 229Table 7.4 The relationship between water depth (m) and total sedimentary netNH4 production and NH

    4 flux across the sedimentwater interface in various marine

    environments. Rates are given as mmol m2 d1. Negative values indicate uptake

    Sediment location DepthNet NH4production

    NH4flux Reference

    Mangrove forest, Thailand 0 1.00 0.40 1Knebel Vig, Denmark 4 2.00 1.00 2W. Kattegat, Denmark 20 3.60 0.90 3Bering/Chukchi Shelf, USA 50 2.10 0.80 4E. Kattegat, Denmark 70 2.80 1.37 3Washington Shelf, USA 90 3.83 0.48 5Washington Shelf, USA 180 0.49 0.10 5Mexican Shelf, Mexico 240 0.35 0.00 6Svalbard Shelf, Norway 300 0.25 0.14 7South Island Shelf, New

    Zealand500 0.62 0.10 8

    Svalbard Shelf, Norway 530 0.06 0.00 7Svalbard Shelf, Norway 1010 0.07 0.00 7Mexican Shelf, Mexico 1020 0.08 0.00 6

    References: 1, Kristensen et al. (2000); 2, Lomstein et al. (1998); 3, Blackburn and Henriksen

    (1983); 4, Henriksen et al. (1993); 5, Christensen et al. (1987); 6, Kristensen et al. (1999); 7,

    Blackburn et al. (1996); 8, Kaspar et al. (1985).When the availability of NH4 to microorganisms for assimilation isconsidered, we must also keep in mind that a fraction of the mineralized

    NH4 becomes reversibly adsorbed onto sediment particles (Mackin andAller, 1984) (Figure 7.12). The equilibrium distribution of NH4 betweenporewater and particles has been successfully described with a linear adsorp-

    tion coeYcient, KS (Rosenfeld, 1979; Mackin and Aller, 1984) (Figure 7.13):

    KS 1 ff rsCads


    where Cads is the concentration of adsorbed NH4 (mmol g

    1 dry solids),Cpw is the concentration of porewater NH

    4 (mmol cm

    3pw), f is porosity

    (cm3pw=cm3total sed), and rs is sediment density (g cm

    3solids). Mackin and Aller

    (1984) have reported a relatively consistent KS value of 1.3 0.1 for a widerange of marine environments, meaning that approximately 57% of the

    produced NH4 is adsorbed to sediment particles while 43% remains indissolved form. Values for KS may be somewhat lower in biogenic and

    very porous sediments. There is the possibility, however, that in some

    cases KS may decrease with increasing NH4 concentrations as adsorption

    sites on the particles become highly loaded (van Raaphorst and Malschaert,


  • 230 CANFIELD ET AL.4.5. New versus regenerated nitrogen in pelagic ecosystems

    Two main sources of nitrogen supply water column primary producers:

    Figure 7.13 Examples of NH4 adsorption from typical anoxic marine sediments.Modified from Mackin and Aller (1984).regenerated nitrogen, originating from the mineralization of organic mat-

    ter in the euphotic zone of the water column, and new nitrogen trans-

    ported to the euphotic zone from elsewhere or fixed there by prokaryotes

    (Dugdale and Goering, 1967). New nitrogen may come from the atmosphere

    as wet and dry deposition, from continental sources as river runoV, from Nfixation or from sediments and the water column below the euphotic zone

    (Figure 7.14). Only new nitrogen, mainly in the form of NO3 , representsnet organic matter production in the system.

    The nitrogen cycle of the Kattegat and the Danish Belt Sea has been

    studied intensively and is used here as an example of coastal marine nitrogen

    cycling (Table 7.5). The annual nitrogen demand by phytoplankton in

    the entire area (31,000 km2) is estimated at 1580 103 t (Richardson,1996). Known sources of new nitrogen, however, account for only

    about 20% of this demand; the remainder is made up from nitrogen regener-

    ation in the water column. The nitrogen budget is constructed from both

    nitrogen sources and sinks. Of the new nitrogen sources, land runoVand Kattegat bottom water are the most important, although there is a

    significant source from atmospheric deposition as well (Table 7.5). Sediment

  • Figure 7.14 Flow diagram showing sources and sinks of nitrogen for primaryproducers in the photic zone of oceans. Heavy arrows represent organic nitrogen,and light arrows represent inorganic nitrogen. Processes are separated into theupper layer (including the photic zone), water masses below the pycnocline,and underlying sediments. Deposition represents dry and wet atmosphericinputs; run-oV represents river, diVuse and sewage discharges from land (import run-oV deposition); export represents the organic nitrogen, which is removedfrom the system via ocean currents; regeneration represents microbial con-version of organic nitrogen to inorganic nitrogen; denitrification represents theloss of combined nitrogen in the form of atmospheric nitrogen; and burial repre-sents nitrogen, which is permanently buried in sediments. Note that nitrogen fixationis not shown.


  • nitrogen upwelled into the photic zone. Some of this remineralized nitrogen

    Table 7.5 Budget for total nitrogen in the Kattegat and Danish Belt Sea


    Nitrification describes the oxidation of NH4 to NO2 and ultimately to

    NO3 . Nitrification links the most reduced and the most oxidizedforms of the nitrogen cycle, and it exerts considerable influence on nitrogenis also exported in bottom currents out of the system.dentrification is by far the most important sink, nearly balancing in-

    puts from land runoff and atmospheric deposition. The nitrogen reminer-

    alized below the photic zone, including sediments, supplies the new

    N sources and sinks 103 t N y1

    Phytoplankton demand 1580Regenerated in the photic zone 1270Total new nitrogen 310

    Atmospheric deposition 50Run-off from land 140From below photic zone 120

    Regeneration below photic zone 144Denitrification in sediments 196Burial in sediments 3.4Exported in bottom currents from system 23

    Data compiled from Blackburn and Henriksen (1983), Hansen et al. (1994), Richardon (1996).dynamics in aquatic environments (Herbert, 1999). Indeed, as a result

    of nitrification (or uptake by other organisms), the primary product of

    nitrogen mineralization, NH4 , rarely occurs at significant concentrationsin oxic environments. The NO3 produced by nitrification below the photiczone of lake and marine water bodies accumulates. Some of this NO3 may betransported upward to the photic zone as a source of new nitrogen for

    primary producers. Some may also be transported downward to the anoxic

    portion of the water body, if one exists, or to the sediments, where the NO3can to be reduced to N2 (denitrification) or NH

    4 (NO

    3 ammonification).

    Because nitrification oxidizes NH4 to NO3 , with a potential sink in denitri-

    fication, it is indirectly responsible for the loss of nitrogen from the system.

    Below, we explore the mechanisms and kinetics of nitrification and

    the factors regulating the rates.

  • THE NITROGEN CYCLE 2335.1. Biochemistry and thermodynamics of nitrification

    The complete oxidation of NH4 toNO3 requires the transfer of eight electrons.

    The first step is the six-electron oxidation of NH4 to NO2 (Equation 7.11)

    accomplished primarily by bacteria of the genera Nitrosomonas and Nitroso-

    spira. These organisms are aerobic chemolithoautotrophs and function best at

    near neutral pH (7 to 8).NH4 1:5O2 ! NO2 H2O 2H; DG 0 272 kJ mol1 7:11

    This overall oxidation reaction proceeds in at least two steps, with hydroxyl-

    amine (NH2OH) as an intermediate. The first step, involving two electrons, is

    the oxidation of NH4 to NH2OH:

    NH4 0:5O2 ! NH2OHH; DG 0 17 kJ mol1 7:12This step is catalyzed by the membrane-associated enzyme ammonium mono-

    oxygenase, and it does not yield biochemically useful energy. Since NH2OH is

    unstable in aqueous solution and rapidly degrades to N2, N2O or NH4 , the

    energy-gaining second step (Equation 7.13) must be coupled closely to the first


    NH2OHO2 ! NO2 H2OH; DG 0 289 kJ mol1 7:13This reaction involves a four-electron transfer and is catalyzed by the periplas-

    mic enzyme hydroxylamine oxidoreductase.

    The next reaction, oxidation of NO2 to NO3 , is primarily accomplished

    by bacteria of the genera Nitrobacter, Nitrococcus Nitrospira, and Nitros-

    pina. While capable of chemolithoautotrophic metabolism, many of these

    organisms (e.g., Nitrobacter) also augment their autotrophic lifestyle with

    heterotrophic metabolism. Oxidation of NO2 to NO3 is a simple two-

    electron transfer with molecular oxygen as the terminal electron acceptor:

    NO2 0:5O2 ! NO3 ; DG 0 76 kJ mol1 7:14

    The membrane-bound enzyme nitrite oxidase catalyzes this process. Since

    intermediate compounds and byproducts other than NH2OH (e.g., N2O)

    have been identified (Kaplan, 1983), it appears that the nitrification processes

    may involve more intermediate steps than depicted above.

    The electron donors for nitrification have high E 00 values for both theNH2OH=NH

    4 couple, which is close to 0 V, and the NO

    3 =NO

    2 couple,

    which is 0.43 V. In either case, these redox couples are too oxidized to reduce

    NAD to NADH (E00 0.32 V), and so a reverse electron transportsystem is required to produce NADH (see Chapter 3). This energy-requiring

  • process limits the amount of ATP produced per NH4 oxidized, reducing thegrowth yield (Kaplan, 1983). Growth yields of nitrifying bacteria are further

    aVected by the energy demands of CO2 fixation. Like many other aerobicchemolithoautotrophs, nitrifiers employ the Calvin cycle (reductive pentose

    phosphate cycle) for CO2 fixation, which requires 3 moles of ATP for every

    CO2 fixed (Chapter 4). Nitrifiers are, therefore, extremely slow growing, and

    a small microbial biomass must oxidize large amounts of reduced nitrogen to

    maintain growth (Figure 7.15). Thus, the thermodynamic efficiency (energy

    conserved for growth: metabolically available energy) is only 14% for

    Nitrosomonas and 310% for Nitrobacter, much less than aerobic hetero-

    trophs (Fenchel and Blackburn, 1979).

    The chemolithoautotrophic oxidation of NH4 might also occur underanoxic conditions. Luther et al. (1997) suggested the possibility of NH4oxidation with MnO2 as an electron acceptor. Two pathways were

    considered, one forming N2 (Equation 7.15), and the other forming NO3

    (Equation 7.16).

    2NH4 3MnO2 4H ! 3Mn2 N2 6H2O; DG0 659 kJ mol17:15

    234 CANFIELD ET AL.NH4 4MnO2 6H ! 4Mn2 NO3 5H2O; DG0 322 kJ mol17:16

    Figure 7.15 Cell yield of various prokaryotes on energy sources with diVerent freeenergies of reaction. O2 is the electron acceptor in all cases. Modified from Madiganet al. (2003).

  • or a combination of both (Figure 7.16). The pathway from reduced nitrogen to

    THE NITROGEN CYCLE 235NO3 may involve intermediates such as NH2OH, NO2 and various organic

    compounds (e.g., amides, aminopropionic acids, nitropropionic acids, nitro-

    soethanol). Heterotrophic nitrifiers gain no energy during nitrification, and it

    has been suggested that intermediates formed have specific functions as micro-

    bial growth factors or as biocidal agents. Heterotrophic nitrification has been

    studied mostly in terrestrial ecosystems, but it is also present in aquatic envir-

    onments (Mevel and Prieur, 1998). However, our knowledge on the role of

    heterotrophic nitrification in aquatic ecosystems is still limited.

    5.2. Phylogeny of chemolithoautotrophic nitrifiers

    The numbers and diversity of organisms identified as chemolithoautotrophic

    nitrifiers are low compared with denitrifiers. The two broad classes of

    organisms, NH4 oxidizers and NO2 oxidizers, are physiologically unrelated,

    and they also employ two very diVerent enzyme systems for the energy-gaining oxidation processes. The phylogenetic relationships of several spe-

    cies of NH4 oxidizers, based on 16S rRNA gene sequences, show twodistinct groups within the proteobacteria. One group, based on the sequence

    of a single NH4 oxidizer, Nitrosococcus oceanus, is deeply branching withinthe -proteobacteria (Woese et al., 1985). The second group contains themajority of cultured strains and forms a tight cluster within the -proteo-bacteria. This group can be subdivided into two clades, corresponding to

    Nitrosomonas spp. and Nitrosospira spp. (Figure 7.17) (Head et al., 1993).

    Within each clade there are a number of closely related strains, some of

    which group into further clusters. Sequences from 16S rDNA amplifiedLuther et al. (1997) further suggested that analogous cycles involving iron

    might be established in ecosystems with low pH, such as many lakes.

    However, Thamdrup and Dalsgaard (2000) could find no evidence for

    anaerobic NH4 oxidation with Mn oxides in Mn oxide-rich marine sedi-ments, where it might be expected, and no direct evidence exists to support

    an analogous cycle with Fe.

    It has long been recognized that many heterotrophic bacteria and fungi can

    oxidize organic nitrogen by nitrification. The organisms responsible for this

    process include Bacteria of the genera Alcaligenes (Papen et al., 1989) and

    Arthrobacter (Brierley and Wood, 2001), as well as certain fungi (Stroo et al.,

    1986). The nitrification rate of heterotrophs is generally 103 to 104 times slower

    than for chemolithoautotrophs. The biochemistry of heterotrophic nitrifica-

    tion is not fully understood, and it appears that diVerent organisms mayemploy unique enzyme pathways, which are also diVerent from those used bychemolithoautotrophs (Jetten et al., 1997). During heterotrophic nitrification,

    nitrogen oxidation can proceed by an inorganic pathway, an organic pathway

  • 236 CANFIELD ET AL.directly from marine sediments and soil reveal at least seven clusters of NH4oxidizers within the -proteobacteria, three within the Nitrosomonas cladeand four within the Nitrosospira clade (Stephen et al., 1998).

    On the basis of ultrastructural properties, four genera of chemolithoauto-

    trophic NO2 oxidizers have been described. Species of the genus Nitrobacterare facultative chemolithoautotrophs and are found in both terrestrial and

    aquatic environments. They can grow aerobically with NO2 as the electrondonor and by the oxidation of simple organic compounds (Kaplan, 1983).

    Members of the genera Nitrococcus, Nitrospina, and Nitrospira are obligate

    chemolithoautotrophs isolated from marine habitats (Watson et al., 1986).

    Based on 16S rRNA sequence data, the Nitrobacter strains constitute a closely

    related subcluster within the a-proteobacteria. Other NO2 oxidizers such asNitrococcus belong to the -proteobacteria, whereas Nitrospina is a member ofthe d-proteobacterial subdivision. Nitrospira represents its own lineage amongthe Bacteria.

    The physiology and phylogenetic distribution of NO2 -oxidizing bacteriain the a and subgroups of the proteobacteria suggest that theyhave descended from photosynthetic purple bacteria. These nitrifiers have

    Figure 7.16 Inorganic and organic pathways of nitrification. Modified fromFocht and Verstraete (1977).

  • THE NITROGEN CYCLE 237Figure 7.17 Phylogenetic tree based upon 16S rDNA sequences from NH4 oxidi-retained the general structural features of the ancestral photosynthetic mem-

    brane complex. NO2 oxidizers as a group are apparently not derived fromone ancestral nitrifying phenotype. Two species within the a-proteobacteria,Bradyrhizobium japonicum and Rhodopseudomonas palustris, are closely

    related to Nitrobacter species (Figure 7.18). The main diVerence is that theformer two fix nitrogen and the latter oxidizes NO2 . B. japonicum formsnitrogen-fixing nodules on leguminous plants, while R. palustris is a free-

    living nitrogen-fixing organism. These diVerences led Orso et al. (1994) tospeculate that these three species evolved from a photosynthetic, nitrogen-

    fixing ancestor at about the time of diVerentiation of land plant families. Oneevolutionary line retained nitrogen fixation and lost the ability to photosyn-

    thesize, while the other evolutionary line acquired the ability to nitrify and

    seemingly lost the ability to fix nitrogen and to photosynthesize.

    5.3. Environmental factors affecting nitrification rates

    Environmental conditions control the location and magnitude of nitrification.

    Nitrifiers depend on the availability of O2 and NH4 , and they are inhibited

    by extremes in pH, sulfide concentration, temperature, salinity, and light.

    zers, Nitrosomonas spp. and Nitrosospira spp., within the -subdivision of proteobacter-ia. Scale bar indicates 1% sequence divergence. Adapted from Kowalchuk et al. (1997).

  • tively low O2 conditions. Thus, the nitrifier Nitrosomonas europaea accumu-

    238 CANFIELD ET AL.lated at low O2 concentrations near the oxicanoxic interface in a stratified

    temperate lake (Voytek and Ward, 1995), and cultures of Nitrosomonas marina

    grow best at O2 concentrations around 5% of air saturation (Goreau et al.,Environmental eVects on nitrification have been studied intensively in labora-tory cultures and by in situ and laboratory measurements on natural materials.

    5.3.1. Substrates

    Because nitrifiers are aerobes, their activity is ultimately constrained by the

    availability of O2, and required minimum concentration of O2 ranges from 1 to

    6mM (Henriksen and Kemp, 1988). Surprisingly, many nitrifiers prefer rela-

    Figure 7.18 Phylogenetic tree based upon 16S rRNA sequences from NO2 oxidi-zers, Nitrobacter spp., within the a-subdivision of proteobacteria. E. coli is shown asthe outgroup for comparison. Scale bar indicates 1% sequence divergence. Adaptedfrom Teske et al. (1994).1980). This is not always the case, however, as Rysgaard et al. (1994) observed

    maximum potential nitrification rates of sediment slurries from a freshwater

    lake at O2 concentrations between 150 and 400mM (Figure 7.19). At higher O2concentrations, the nitrification rate was somewhat inhibited. Overall, the

    composition of the nitrifying community, and its O2 optimum, appears to be

    location specific and depends on adaptations to the prevailing environmental


    Indeed, diVerent environmental adaptations can be found among closelyrelated nitrifying species. Thus, the two Nitrosomonas species N. eutropha

    and N. europaea are closely related by 16S rDNA sequence analysis, yet

    N. eutropha dominates in oxygen-poor profundal sediments of an oligo-

    trophic lake, while N. europaea primarily inhabits sediments from the

    oxygen-rich littoral zone (Whitby et al., 1999).

    Nitrifiers in sediments may survive periods of inactivity when they are

    periodically or persistently exposed to anoxia, and they recover their activity

  • THE NITROGEN CYCLE 239Figure 7.19 Potential nitrification in sediment from a freshwater lake as a func-tion of O2 concentrations in the overlying water. The equation describing O2 as botha substrate (Michaelis-Menten kinetics) and a non-competitive inhibitor is shown.Values for the constants, Rmax, Ki and KS, which provide the best fit to the datapoints, are also given. Modified from Rysgaard et al. (1994).instantly following O2 exposure (Henriksen et al., 1981). The physiological

    basis of this tolerance to anoxia is not well understood. However, a number

    of both NH4 and NO2 oxidizers are apparently capable of heterotrophic

    growth using low-molecular organic substrates as electron donors while

    performing partial or complete dissimilatory nitrate reduction. Nitrifiers in

    marine sediments, by contrast, are almost completely inactivated by expo-

    sure to free sulfide (Kaplan, 1983), and they may require days to weeks to

    recover completely (Joye and Hollibaugh, 1995). Thus, the ability of nitri-

    fiers to recover from environmental O2 fluctuations may be impeded if

    sulfide exposure also occurs.

    Ammonium oxidation is generally the rate-limiting step of nitrification

    when there are ample supplies of O2 and when NO2 accumulation is unlikely

    (Kaplan, 1983). However, since NH4 oxidizers generally have higher aYnityfor O2 than NO

    2 oxidizers, accumulation of NO

    2 and N2O may occur at

    very low O2 concentrations (Henriksen and Kemp, 1988).

    Generally, the oxidation of NH4 and NO2 by nitrifiers follows Michaelis-

    Menten kinetics, and a range of Km values have been observed. The nitrifi-

    cation of NH4 to NO2 in eutrophic environments has Km values in the

    range of 50700mM NH4 , and Km values of 350600mM NO2 are

    observed for NO2 oxidation to NO3 (Focht and Verstraete, 1977; Koops

  • and Pommerening-Roser, 2001). However, nitrifiers from oligotrophic open

    ocean and lake environments are adapted to much lower substrate concen-

    trations, and Km values ranging from 0.1 to 5 mM have been observed forboth NH4 and NO

    2 oxidation (Olson, 1981a; Koops and Pommerening-

    Roser, 2001). Thus, the substrate aYnity of nitrifying populations adapts tothe substrate availability in the environment. Marine populations of nitri-

    fiers are rather insensitive to perturbations in NH4 and NO2 concentra-

    tions, as their activity is usually saturated at normal environmental levels of

    NH4 and NO2 (Ward, 2000). However, intense competition with phyto-

    plankton or benthic microalgae for NH4 and NO2 can limit the growth of

    nitrifiers in the upper photic zone of oligotrophic waters and in surface

    sediments (Jensen et al., 1994). Nitrification rates in eutrophic lakes, rivers,

    and coastal marine areas are generally higher than in coastal oligotrophic

    ocean environments (Table 7.6). Not surprising, eutrophic environments

    receiving elevated supplies of NH4 from organic compound mineralizationand anthropogenic input support denser and more active nitrifier popula-

    tions. The activity of nitrifiers may be lower in rivers and other environments

    receiving reactive organic matter poor in nitrogen. In this case, the nitrifiers

    must compete with heterotrophic bacteria for the available NH4 (Strausset al., 2002).

    240 CANFIELD ET AL.Table 7.6 Range of depth-integrated nitrification and denitrification rates insediment (mmol N m2 d1), and volume-specific rates in the water column(mmol Nl 1 d1), from various aquatic ecosystems

    Ecosystem Nitrification Denitrification

    LakeSediment 220 0.26Water 01 0.22a

    RiverSediment 323 0.920Water 04 02a

    Coastal marine (100 m)Sediment 0.0030.1 0.0030.6Water 0.0010.01 0.0010.01a

    aOnly valid in the rare cases when the water column or part of it is devoid of oxygen.

    Data compiled from Kaplan (1983), Christensen and Rowe (1984), Koike and Srensen (1988),

    Seitzinger (1988), Jensen et al. (1993), Omnes et al. (1996), Ward (1996), Sjodin et al. (1997),

    Iriarte et al. (1998), Lorenzen et al. (1998), Herbert (1999), and Pauer and Auer (2000).

  • 5.3.2. pH and surface growth

    The growth and activity of most chemolithoautotrophic nitrifiers are opti-

    mal in the neutral to slightly alkaline pH range (pH 7 to 8.5) (Focht and

    Verstraete, 1977). The restricted optimal pH range for nitrification is mainly

    due to the toxicity of free NH3 at high pH and of nitrous acid (HNO2) at low

    pH. Nitrobacter spp. are sensitive to both high and low pH levels, whereas

    Nitrosomonas spp. are more aVected by alkaline conditions when high NH3concentrations are present (Figure 7.20).

    Apart from acid lakes, most natural aqueous environments have pH levels

    within the optimal range for nitrifiers. Sedimentary environments, on the

    other hand, can experience pH extremes well outside of the optimal range.

    For example, shallow water sediments with benthic microalgae may reach

    pH 10 in the photic zone during illumination (Revsbech et al., 1983). Also,

    sulfide-rich salt marsh and mangrove sediments, with intensive sulfide oxi-

    dation, may have pH

  • identified in active compost at 53 8C (Pel et al., 1997). Autotrophic nitrificationwas not found in this mixed compost population suggesting that thermophilic

    242 CANFIELD ET AL.chemolithoautotroph nitrification apparently is rare in nature.

    When metabolizing within their tolerance range, pure cultures and natural

    populations of nitrifiers respond exponentially to temperature, with Q10values ranging from 2 to 3 (Pomeroy and Wiebe, 2001). In nature, nitrifying

    communities tend to adapt to the temperature of the environment. Thus,

    nitrifying populations in arctic sediment from Svalbard, Norway have a

    temperature optimum of only 14 8C, well within the psychrophilic range.By contrast, the temperature optimum was near 40 8C for nitrifiers fromwarmer temperate sediments oV Germany (Thamdrup and Fleischer 1998)(see also Chapter 2 for more discussion on temperature adaptations). In

    some instances, individual populations can also adapt their metabolic range

    to the temperature of the environment in which they live. For example, an

    NH4 -oxidizing strain was isolated from arctic waters with growth minimumof 5 8C, an optimum of 22 8C, and a maximum growth temperature ofabout 29 8C when adapted to 5 8C. Cells grown at 25 8C, on the other hand,showed an optimum temperature of 30 8C and a maximum of about 38 8C.inhibited completely under these extremes of pH sometimes found in sedi-

    ments. Thus, species of Nitrosospira and Nitrosovibrio can be active in acid

    soils and sediments at pH values around 4 (de Boer and Laanbroek, 1989),

    while Nitrobacter alkalicus has been isolated from sediments of soda lakes

    with pH values up to around 10 (Sorokin et al., 1998). Heterotrophic

    nitrifiers may also contribute to NO3 formation under acidic conditions(Brierley and Wood, 2001).

    Sediments generally host, on a per volume basis, nitrifier populations

    orders of magnitude greater than in the water column Thus, nitrifiers

    are typically present in the range of 105 to 107 cells cm3 in sediments,while in the water column they range from 101 to 104 cells cm3 (Fochtand Verstraete, 1977; Pauer and Auer, 2000). The strong association of

    nitrifiers with the particulate- and organic-rich fractions of sediments is

    linked to the high availability of NH4 derived from microbial nitrogenmineralization of the organic particles. Furthermore, the pH-buffering of

    clay and silt particles is likely to help retain an optimal pH for the nitrifiers.

    5.3.3. Temperature

    The temperature optimum for typical mesophilic nitrifiers in pure culture

    ranges from 25 to 35 8C, and growth usually occurs from 3 to 45 8C (Herbert,1999). Thermophilic heterotrophic nitrifiers with temperature optima of

    around 65 8C have also been isolated from deep-sea hydrothermal vents(Mevel and Prieur, 1998). Heterotrophic thermophilic nitrifiers have been

  • Nitrification rates in these warm-adapted populations exceeded the

    cold-adapted cells only above 25 8C (Jones and Morita, 1985).Nitrifying bacteria in shallow-water sediments from temperate regions are

    subject to large seasonal changes in temperature, and they exhibit the highest

    cell-specific activity during the warm summer months. However, when in situ

    measurements of nitrification in sediments are compared on a seasonal basis,

    diVerent patterns emerge. Distinct summer maxima for nitrification havebeen observed, for example, in Aarhus Bay, Denmark (Figure 7.21) (Hansen

    et al., 1981), the Tay Estuary, Scotland (Macfarlane and Herbert, 1984), and

    the middle reaches of Narragansett Bay, USA (Seitzinger et al., 1984).

    Nitrification rates in some shallow Danish fjords (Figure 7.21) (Hansen

    et al., 1981) and the Providence River station in Narragansett Bay (Jenkins

    and Kemp, 1984), on the other hand, are lowest during the summer and

    highest during the winter. The reduced nitrification activity in the summer

    has been attributed primarily to reduced availability of O2 to the nitrifying

    population. The warm conditions during summer reduce O2 solubility, and

    high rates of benthic respiration increase O2 demand. In concert, both of

    these factors reduce the depth of O2 penetration into the sediment. Other

    factors contributing to reduced rates of summer nitrification include greater

    competition for NH4 by heterotrophic bacteria and sulfide toxicity.

    THE NITROGEN CYCLE 243Figure 7.21 Seasonal variations in nitrification for a sediment from the estuary,Norsminde Fjord, Denmark, and at 17m depth from Arhus Bay, Denmark. Watertemperature is shown as the dotted line. Modified from Hansen et al. (1981).

  • 5.3.4. Salinity

    Nitrifiers are adapted to the salinity prevailing in their environment, al-

    though individual species are generally able to acclimate to a broad salinity

    range. The tolerance to varying salinities, however, diVers from species tospecies. Most freshwater species rapidly reduce their activity when salinity

    increases, and some do not grow at all in sea water, whereas the opposite

    pattern is observed for many marine species. NH4 oxidizers from estuarineenvironments have optimum activity at intermediate salinities (510), and

    many of them can grow in the entire salinity range of 035. For example,

    Nitrosomonas eutropha, isolated from the Elbe River, is an euryhaline species

    exhibiting at least 75% of the maximum activity in the whole of the salinity

    from 0 to 30 (Stehr et al., 1995) (Figure 7.22). Other nitrifiers isolated from

    the same river exhibit diVerent salinity responses. Thus, N. oligotrophaappears to be a freshwater species, which is completely inhibited at salinities

    above 10, and N. europaea has a higher salinity tolerance but cannot func-

    tion at full seawater salinity. In the Schelde estuary, changes in salinity along

    the estuary cause community shifts of the NH4 oxidizing population (de Bieet al., 2001). The most dramatic change occurs in the estuarine region with

    the sharpest salinity gradient.

    244 CANFIELD ET AL.Figure 7.22 Sensitivity to increasing salinity of three species of NH4 -oxidizingNitrosomonas, isolated from the lower River Elbe, Germany. Modified from Stehret al. (1995).

  • 5.3.5. Light

    variety of 15N tracer methods has been proposed. The most promising of these

    is the relatively simple, accurate, and versatile nitrogen isotope pairing tech-

    THE NITROGEN CYCLE 245nique of Nielsen (1992), which provides simultaneous measurements of both

    nitrification and denitrification (see text box below).

    The use of newly developed microelectrodes can overcome biases resulting

    from uneven dispersal of inhibitors or tracers, or unwanted side eVects from

    inhibitors. Microelectrodes with very high spatial resolution can measure the

    fine-scale distributions of O2 and NO3 in sediments (Figure 7.23). At steady

    state, reaction diVusion models can be used to estimate rates of nitrification

    and denitrification and to determine the location of the processes in relation to

    the chemical profiles (Revsbech and Jrgensen, 1986).Light often inhibits nitrification in the open ocean and in lake environ-

    ments (Vanzella et al., 1989; Ward, 2000). Although NO2 oxidizersare usually considered more sensitive to sunlight than NH4 oxidizers,this generalization may be obscured by species-specific as well as dose-

    and wavelength-dependent responses (Guerrero and Jones, 1996). However,

    the higher sensitivity of NO2 oxidizers to light supports the hypothesisof Olson (1981b) that light inhibition is responsible for the commonly ob-

    served near-surface NO2 maximum in the water column of many oceanicenvironments.

    Measuring nitrification and denitrification

    Specific inhibitors are widely used to determine nitrification and denitrifica-

    tion rates. In using these, the process of interest is blocked and the accumula-

    tion of unused substrates is determined (e.g., NH4 or NO2 for nitrification and

    N2O for denitrification). The most commonly used inhibitors of nitrification

    are nitrapyrin, allylthiourea, and chlorate (Henriksen and Kemp, 1988). The

    former two inhibit NH4 oxidation and the latter inhibits NO2 oxidation.

    Inhibitor-based quantification of denitrification is usually done by acetylene

    inhibition (Srensen, 1978). Acetylene inhibits N2O reductase, with a resulting

    accumulation of N2O. Problems sometimes exist with these inhibitor-based

    methods, however, limiting the usefulness of the results. For example, sediment

    denitrification rates obtained with the acetylene-block technique may be

    too low because coupled nitrification-denitrification is excluded due to the

    simultaneous inhibition of nitrification by acetylene.

    Alternatively, nitrification can be estimated from inorganic nitrogen

    fluxes (Christensen and Rowe, 1984), whereas denitrification can be measured

    directly as the production of N2 (Seitzinger, 1993). These methods, however,

    lack accuracy, and the experimental design can be elaborate. More recently, a

  • 246 CANFIELD ET AL.Dissimilatory NO3 reduction is a microbial process in which, NO3 is

    reduced with various electron donors by an energy-gaining metabolism in6. DISSIMILATORY NITRATE REDUCTION

    Figure 7.23 (A) Steady-state concentration profile of NO3 in a freshwater lakesediment measured with a NO3 microsensor. A concentration of 300 mM of NH

    4 was

    added to the overlying water. (B) The corresponding distribution and magnitude of thevolume-specific nitrification (Nitr.) and NO3 reduction rates (NO

    3 red.) are indicated

    by the black and grey areas, respectively. Modified from Jensen et al. (1993).the absence or near absence of O2. NO2 is the first intermediate in this

    reduction, and based on the fate of NO2 , three diVerent pathways can bedistinguished (Figure 7.3) (Bonin et al., 1998; Thamdrup and Dalsgaard,

    2002): (1) reduction to gaseous products (N2O or N2) by denitrification,

    (2) reduction to NH4 in a process termed NO3 ammonification or dissimi-

    latory NO3 reduction to ammonium, and (3) reduction to N2 coupledto the oxidation of NH4 by the newly discovered anammox process.Most dissimilatory NO3-reducing prokaryotes are facultative anaerobic

    heterotrophs utilizing either dissolved low-molecular-weight carbon sources

    (e.g., Pseudomonas spp.) or one-carbon compounds (e.g., Alcaligenes spp.)

    (Table 7.7). Other dissimilatory NO3 reducers grow as chemolitho-autotrophs by oxidizing reduced inorganic compounds, such as H2 (Para-

    coccus spp.), H2S (Thiobacillus spp.). Also, the anammox bacteria of the

    Planctomycetales are chemolithoautotrophs.

    NO3 reduction in sediments occurs below the oxic surface zone. In openwaters, NO3 reduction may occur in O2-depleted zones of the oceans(oxygen minimum zones) and in the anoxic hypolimnion of lakes when

    NO3 is present. River plumes rich in suspended particles support NO3

    reduction within aggregates, where anoxic microenvironments occur in an

    otherwise oxygenated water body (Omnes et al., 1996).

  • THE NITROGEN CYCLE 247The isotope pairing technique

    The nitrogen isotope pairing technique (Nielsen, 1992) was developed to

    measure nitrification and denitrification rates in sediments but can be applied

    to other environments as well. The water in the enclosed experimental system is

    enriched with 15NO3 , which mixes with the14NO3 of the naturally occurring

    NO3 . The formation of single-labeled (14N15N) and double-labeled (15N15N)

    dinitrogen pairs by denitrification is measured by mass spectrometry. Denitri-

    fication, including the formation of unlabeled (14N14N) dinitrogen, can be

    determined assuming random isotope pairing by denitrification of the uniform-

    ly mixed NO3 species. Rates of denitrification based on15NO3 (D15) are

    calculated as follows:

    D15 14N15N 215N15N

    Because the production of 14N14N cannot be precisely measured due to the

    large natural background, denitrification based on 14NO3 (D14) is calculatedaccording to the following:

    D14 14N15N=215N15ND15D14 thus represents the indigenous denitrification rate of the system.

    To gain further insight into the sources of NO3 , denitrification may bedivided between the activity based on NO3 from the overlying water (D


    and that based on NO3 from nitrification (Dn14):

    Dw14 D15=e

    Dn14 D14 Dw14where e is the 15N-labeled fraction of the NO3 in the surrounding water(Figure 7.24).

    NO3 ammonification (DNRA) can be estimated as

    DNRA FNH4 x Dy=nwhere FNH4 is the formation of NH

    4 , Dy is the change in

    15N labeled fraction

    of NH4 , and n is the15N-labeled fraction of NO3 that is reduced to NH


    (modified from Rysgaard et al., 1993).

    Total nitrification (Nt) can be determined as

    Nt Dn14 Nfwhere Nf is the measured net formation of unlabeled NO

    3 . The fundamental

    limitation of the isotope pairing technique is the demand for a uniform mixing

    of the added 15NO3 with the endogenous sources of14NO3 . The assumption of

    random isotope pairing is not valid when anammox contributes significantly to

    N2 production, and further measurements are needed to determine denitrifica-

    tion rates.

  • Table 7.7 Examples of archaeal and bacterial genera harboring denitrifyingspecies with the source of energy, electrons, and carbon indicated

    Genus Energy/electron source Carbon source

    ArchaeaHalobacterium Chemoorganotrophic HeterotrophicPyrobaculum Chemoorganotrophic Heterotrophic

    BacteriaPseudomonas Chemoorganotrophic HeterotrophicAquaspirillum Chemoorganotrophic HeterotrophicAzospirillum Chemoorganotrophic HeterotrophicPseudomonas Chemoorganotrophic HeterotrophicAlcaligenes Chemoorganotrophic HeterotrophicRhodobacter Photolithotrophic AutotrophicBeggiatoa Chemolithotrophic AutotrophicThiobacillus Chemolithotrophic AutotrophicThioploca Chemolithotrophic AutotrophicParacoccus Chemolithotrophic Autotrophic

    Figure 7.24 Rates of nitrification and denitrification in stream sediment asdetermined by the isotope pairing technique. Black bars (Nitr.) indicate couplednitrification-denitrification (equivalent to nitrification when no NO3 eZux occursacross the sedimentwater interface) and white bars indicate denitrification of NO3diVusing from the overlying water (Denitr.). Scenarios are shown with two diVerentNO3 concentrations and with the addition of the nitrification inhibitor (n.i.)thiourea. Modified from Nielsen (1992).


  • THE NITROGEN CYCLE 249Respiratory NO3 reductases are membrane-bound complexes consisting of

    two to three subunits (Zumft, 1997), and they are remarkably similar among

    denitrifiers in subunit structure and molecular weight. The enzyme contains

    eight to twelve Fe-S groups and one atom of molybdenum in the catalytic

    subunit. The concentration of NO3 reductase in the cytoplasmic membraneof denitrifiers may be as high as 25% of the total membrane proteins

    (Stouthamer, 1988). The mechanism of coupled proton-electron transfer

    during NO3 reduction involves the interaction of Mo(IV) with NO3 .

    A fully protonated ligand (X) is attached to the Mo atom. As the Mo(IV)

    reduces NO3 by two electrons, it is oxidized to Mo(VI) and the donor ligandX transfers a proton to NO3 , which then splits into NO

    2 and hydroxide:

    HXMoIV NO3 ! HXMoIV NO3 !XMoVI2 NO2 OH 7:18

    The Mo(VI) is reduced again by the Fe-S clusters in the enzyme (Stouthamer,


    6.1.2. NO2 reductase

    Two main types of respiratory NO2 reductases have been isolated from deni-trifiers: a tetraheme (cytochrome cd1-Nir) enzyme and a copper-containing

    (CuNir) enzyme (Zumft, 1997). About 75% of the known denitrifying strains

    have the cd1-Nir enzyme, and the two enzymes are not found together in the

    same strain (Gamble et al., 1977). The cd1-Nir enzyme consists of two

    identical subunits, each containing two types of prosthetic groups: heme c,

    which is covalently linked to the protein, and a non-covalently bound

    chlorine type, heme d1. Electrons enter cd1-Nir at heme c (Fe2) and are

    transferred to NO2 via heme d1. Transfer of one electron is associated withprotonation and removal of water and yields one molecule of NO:6.1. Biochemistry of denitrification

    Denitrification is a major sink in the nitrogen cycle, converting NO3 to N2and removing fixed nitrogen from the environment. It consists of a number

    of respiratory reduction steps (Equation 7.17).

    NO3 ! NO2 ! NO ! N2O ! N2 7:17The characteristics and function of the enzymes (reductases) catalyzing each

    of the reduction steps are briefly discussed in the following sections.

    6.1.1. NO3 reductase

  • As NO is highly reactive with O and transition metals. It is toxic to living

    2 3

    250 CANFIELD ET AL.2NorB-Fe 2NO 2H ! 2NorB-Fe N2OH2O 7:21However, the exact mechanisms of NO reduction are still not fully understood.

    6.1.4. N2O reductase

    The conversion of N2O to N2 is the last step of the denitrification pathway.

    The role of N2O as an obligatory intermediate is confirmed by the fact that

    many denitrifying bacteria are able to grow at the expense of N2O as the sole

    electron acceptor. N2O reductase (NosZ proteins) consists of two subunits

    with about four Cu atoms per subunit. Parallel pathways of electron trans-

    fer, and sometimes alternative electron donors for the enzyme, exist in

    diVerent denitrifiers, but there are strong indications for the involvementof c- and b-type cytochromes:

    2NosZ-Cu N2O 2H ! 2NosZ-Cu2 N2 H2O 7:222

    organisms and rarely accumulates in cells, but is either excreted to the

    surroundings or rapidly reduced intracellularly.

    6.1.3. NO reductase

    NO reductase, which reduces NO to N2O, was the last identified enzyme

    involved in denitrification. The enzyme is composed of two subunits, NorB

    (highly hydrophobic heme b-type cytochrome) and NorC (membrane-bound

    monoheme c-type cytochrome) (Zumft, 1997). Electrons are believed to

    enter the iron-containing reaction center in the NorB subunit via the NorC

    subunit and the process proceeds according to the following:d1Fe2 NO2 2H ! d1Fe2 NO2 2H !d1Fe3 NOH2O 7:19

    The product of the cd1-Nir pathway is always NO.

    The Cu-Nir enzyme contains two identical subunits with one copper

    atom each. The transfer of electrons from Cu to NO2 is associated withprotonation and removal of water according to the following:

    Cu NO2 2H ! Cu NO2 2H !Cu2 NOH2O 7:20

    The product of the Cu-Nir pathway is mostly NO, but N2O may be formed

    under strongly reducing conditions and at high pH.

  • 6.2. Biochemistry of NO3 ammonification

    THE NITROGEN CYCLE 251NO3 ammonification, or dissimilatory nitrate reduction to ammonium, isused by organisms to detoxify NO2 , and in some cases as an electron sinkduring fermentation (see also Section 4 in Chapter 9). Some NO3 ammoni-fiers are also true respirers (Welsh et al., 2001). NO3 ammonification istherefore defined as the dissimilatory transformation of nitrogen oxides to

    NH4 . In some cases, but not all, this process is coupled to energy conservation.The complete reduction of NO3 to NH

    4 proceeds in two steps:

    NO3 ! NO2 ! NH4 7:23NO3 ammonification can occur under the same environmental conditions asdenitrification, and because N2 gas is not produced, this process keeps nitro-

    gen in the ecosystem. NO3 ammonification gains the most prominence inhighly reducing environments, particularly in the presence of free sulfide

    (Brunet and Garcia-Gil, 1996).

    Both facultative and obligate anaerobes mediate NO3 ammonification,but the pathways are often diVerent and complex (Bonin, 1996). Since thefirst step in both denitrification and NO3 ammonification is the respiratorytransformation of NO3 to NO

    2 , catalyzed by NO

    3 reductase, these two

    processes separate biochemically only after NO2 is formed.6.1.5. The entire denitrification process

    NO3 reductase and NO reductase are membrane-bound enzymes in mostdenitrifiers, whereas NO2 reductase and N2O reductase are located in the cellperiplasm. Although complete denitrification of NO3 to N2 is most common,a number of denitrifiers lack one or more of the reductases. For example,

    N2O reductase is absent in some Rhizobium and Pseudomonas strains. In some

    cases, organisms that normally contain the full complement of reductase

    enzymes can miss one of them in defective mutants (Stouthamer, 1988).

    Alternatively, the environment may be deficient in one or more trace element,

    such as the metals Fe, Cu, and Mo, which are essential for the biosynthesis of

    denitrification enzymes. Since N2O reductase is only fully operative around

    pH 7 and is inhibited completely at pH 5 due to conformational changes of the

    protein (Zumft, 1997), the end product of denitrification will gradually shift

    from N2 to N2O when pH decreases in the environment.

    Complete denitrification can be viewed as the modular assemblage of four

    partly independent respiratory processes (Zumft, 1997). Complete denitrifica-

    tion is achieved only when all four modules are activated simultaneously

    (Figure 7.25). When there is no overlap between the modules, or only pairwise

    overlap, accumulation of intermediates such asNO2 , NO, and N2O mayoccur.

  • 252 CANFIELD ET AL.NO3 ammonification is widespread among heterotrophic and chemoau-totrophic Bacteria, and the reaction can be considered a short circuit in the

    nitrogen cycle (Figure 7.3). A number of diVerent reductases are known tocatalyze the terminal six-electron reduction of NO2 to NH

    4 . Common to

    them all is the presence of heme, which has iron and sulfur clusters in the

    oxidation-reduction centers. A NO2 reductase containing six c-type hemegroups per molecule has been purified from Desulfovibrio desulfuricans, and

    it reduces NO2 by the following reaction (Liu and Peck, 1981):

    6c-Fe2 NO2 8H ! 6c-Fe3 NH4 2H2O 7:24Some of the most common NO3 ammonifiers in coastal marine sediments

    are fermenting bacteria within the genera Aeromonas and Vibrio. These

    organisms couple the reduction of NO2 to the oxidation of NADH pro-duced during glycolysis (Bonin, 1996). We can follow this path as follows:

    pyruvate is generated together with NADH in the first step of glucose


    Figure 7.25 Modular organization of denitrification. Four modules representingthe respiratory systems utilizing NO3 (upper left), NO

    2 (upper right), NO (lower

    right), and N2O (lower left) constitute the overall process. Complete denitrification isachieved only when all four modules are active (zone bordered by heavy lines).Partial overlap of two or three modules occurs naturally in individual denitrifiersand in microbial communities. Modified after Zumft (1997).


    The acetate pathway can only proceed when reactions other than the ethanol

    pathway are available to oxidize NADH and replenish NAD+ in the cells.

    This is where NO2 comes in, as it can act as an alternative electron sink forNADH:

    NO2 NADH ! NH4 NAD 7:27The overall outcome of glucose fermentation by NO3 ammonifiers depends,therefore, on the availability of NO2 ; acetate formation occurs and y/x >0 when NO2 is present, whereas this pathway is inhibited and y/xapproaches 0 when NO2 is deficient.

    Chemolithoautotrophic NO3 ammonification coupled to sulfide oxida-tion has been discovered in giant marine bacteria belonging to the genus

    Thioploca (Schulz and Jrgensen, 2001). Thioploca (see Chapter 9 for further

    discussion) is capable of accumulating up to 500 mM NO3 in its largecentral vacuole from sea water containing 25 mM NO3 (Fossing et al.,1995). The nitrate is used as an electron acceptor in oxidizing sulfide to

    sulfate, with the production of NH4 according to Equation 7.28 (Otte et al.,1999). Other phylogenetically related NO3 accumulating sulfide oxidizers ofthe genera Beggiatoa and Thiomargarita may also reduce NO3 to NH


    during sulfide oxidation, but this still needs to be established.

    HS NO3 H2OH ! SO24 NH4 7:28It appears that both bicarbonate and acetate may act as carbon sources,

    indicating that Thioploca is a facultative NO3 ammonifying chemolithoau-totroph capable of mixotrophic growth.

    6.3. Phylogeny and detection of denitrifiers

    Nearly 130 denitrifying prokaroyte species have been isolated within more

    than 50 genera, and most of these are found within the a, , and sub-divisions of the proteobacteria (Zumft, 1997). Denitrification is, however,glucose 2 NAD ! 2 pyruvate 2 NADH 7:25Among other products, a fraction x of the pyruvate will be fermented

    to ethanol with the consumption of NADH, while a fraction y will be

    fermented to acetate (White, 1995):

  • widespread among the Bacteria, including gram-positive Bacillus spp. and

    the deep-branching Aquifex pyrophilus. Among the Archaea, denitrification

    is described from the halophilic Haloarcula denitrificans, while other halo-

    philes and hyperthermophiles such as Pyrobaculum aerophilum denitrify

    NO2 to N2O.Key enzymes in the denitrification process, NO2 reductase and

    N2O reductase, have been the focus of phylogenetic studies (Hallin and

    Lindgren, 1999; Scala and Kerkhof, 1999). The two structurally diVerentbut functionally equivalent enzymes catalyzing NO2 reduction, the tetra-heme (cd1-Nir) enzyme and the copper-containing (Cu-Nir) enzyme, are

    encoded by the genes nirS and nirK. The two types of nir genes are mutually

    exclusive in a given strain, although the nir type may diVer within the samegenera and even within the same species (Coyne et al., 1989). There is high

    diversity and near-random distribution of the two nir genes among denitri-

    fying genera (Figure 7.26). For example, the genera Alcaligenes and Pseudo-

    254 CANFIELD ET AL.monas contain both nirS and nirK genes. The same nir-type gene among

    otherwise phylogenetically diVerent groups, and the occasional occurrenceof diVerent nir types within the same species, could be an indication ofhorizontal gene transfer (Braker et al., 1998).

    The gene encoding N2O reductase (nosZ) is largely unique to denitrifying

    bacteria and has been used to indicate the presence of denitrifiers in the

    environment (Scala and Kerkhof, 1999). Phylogenies constructed from

    nosZ gene sequences can also be compared with phylogenies from 16S rRNA

    to check for congruent evolution between the two genes. Indeed, comparisons

    show some inconsistencies. For example, the 16S rRNA unrelated Pseudomo-

    nas denitrificans and Paracoccus pantotrophus (formerly Thiosphaera pantotro-

    pha) have almost identical nosZ genes (Figure 7.27). By contrast, the 16S rRNA

    Figure 7.26 Neighbor-joining phylogenetic tree based upon nir gene sequencesfrom denitrifying prokaryotes. Scale bar indicates 2% sequence divergence. Adaptedfrom Braker et al. (1998, 2000).

  • THE NITROGEN CYCLE 255closely related Paracoccus pantotrophus and Paracoccus denitrificans have only

    distantly related nosZ genes. Taken together, these results imply that P. panto-

    trophus has either acquired a 16S rRNA gene from Paracoccus denitrificans (or

    a related organism) or eliminated a functional nosZ gene in favor of a nosZ gene

    from Pseudomonas denitrificans.

    There is a substantial phylogenetic diversity among denitrifiers in both

    freshwater and marine environments. Many of the variety of clones and

    strains obtained by sequence analysis of nir and nos genes from environ-

    mental samples are not represented in culture collections, suggesting that

    denitrification is more widespread among prokaryotes than previously

    anticipated (Scala and Kerkhof, 1999; Braker et al., 2000). Denitrifying

    clones often group according to the environment from which they are

    obtained. For example, there are strong geographic diVerences among deni-trifier populations in marine sediments (Scala and Kerkhof, 1999), and these

    are greater than the diVerence in population structure one might observewith depth in a sediment at a given location. Thus, mixing by burrowing

    animals and other advective mechanisms may limit the gradients in sedimen-

    Figure 7.27 Phylogenetic tree based on nosZ gene sequences from diVerent spe-cies of denitrifiers. Scale bar indicates 10% sequence divergence. Adapted from Scalaand Kerkhof (1999).tary population composition (Braker et al., 2001).

    On the other hand, changes in key environmental parameters within rela-

    tively short geographical distances exert site-specific selection pressure and

    could cause diversification among denitrifiers. Hence, denitrifiers are appar-

    ently well adapted to the environment where they are found. As an example,

    the relatively high input of refractory organic matter into near-coastal sedi-

    ments oV the Washington coast seems to support distinct denitrifier popula-tions compared to continental shelf sediments receiving a greater input of

    labile carbon (Braker et al., 2000). The long-term geographical separation of

    similar environments may also result in species diversification. Thus, enclosed

    basins such as the Baltic Sea and the Black Sea, with anoxic deep water, create

    an ideal environment for denitrifiers at the oxicanoxic interface. However,

    many of the denitrifier species in these two similar environments are

  • There are two major sources of NO3 for denitrifying communities in

    aquatic sediments: (1) NO diVusing into the sediment from the overlying

    due to a close coupling between nitrification in the oxic zone and denitrifica-

    tion in the anoxic zone (Jenkins and Kemp, 1984). The NO produced in the

    256 CANFIELD ET AL.3

    oxic zone will diVuse both into the overlying water and into the anoxicsediment to be denitrified. Even with an oxygen penetration of just 1 mm,3

    water and (2) NO3 produced within the sediment by nitrification. Rates ofdenitrification, and the partitioning between the two sources of NO3 , arestrongly dependent on the NO3 concentration in the overlying water as wellas the penetration depth of O2 into the sediment (Figure 7.28). Since denitri-

    fication is restricted to a thin anoxic layer immediately under the oxic zone, it

    is primarily the thickness of the oxic zone that controls the diVusional supplyof NO3 from the overlying water to the denitrifiers below (Nielsen et al.,1990). The longer the diVusion path for NO3 through the oxic zone, the lesssteep the NO3 gradient and the lower the diVusional supply. Often, howev-er, NO3 generated within the sediment is the main source for denitrificationphylogenically distinct but in some cases closely related, such as Shewanella

    baltica in the Baltic and S. oneidensis in the Black Sea (Brettar et al., 2001).

    6.4. Environmental factors affecting denitrification

    Denitrification is an inducible process, occurring only when O2 is absent or

    nearly absent and when NO3 is present. Factors other than absence of O2,however, may influence denitrification rates in aquatic environments;

    these include temperature and substrate availability (NO3 and organicmatter). The following discussion concerns the principal factors controlling

    denitrification rates in the environment and how they are interrelated.

    We primarily focus on sedimentary environments, as denitrification is a

    ubiquitous and important process in most aquatic sediments (Table 7.6).

    6.4.1. Dependence of NO3 and O2

    Rates of denitrification in cultures and in the environment follow Michaelis-

    Menten kinetics with respect to NO3 (Seitzinger, 1990). In sediments, half-saturation constants (Km) are generally between 2 and 170 mM NO

    3 with an

    average of about 50 mM NO3 . However, values higher than 500 mM NO3

    also have been reported (Joye et al., 1996; Garcia-Ruiz et al., 1998). Accord-

    ingly, denitrification in most natural and unpolluted aquatic environments

    responds to NO3 concentration because concentrations are usually lowerthan 20 mM in these environments. Adequate supplies of NO3 are thereforeessential to maintain the denitrification process at high rates.

  • oxygen minimum zones requires O2 concentrations 3 mM (Codispoti et al.,

    THE NITROGEN CYCLE 2572001), while somewhat higher limiting concentrations have been inferred for

    other waters and sediments (Seitzinger, 1988). In a few unusual cases, as with

    the denitrifying mixotroph, Paracoccus (Thiosphaera) pantotrophus, NO3and O2 can be used simultaneously as electron acceptors at O2 concentra-

    tions up to 90% of air saturation (Robertson and Kuenen, 1984). Thus, in

    some instances, active denitrification occurs under well-oxygenated condi-

    tions (Robertson et al., 1995). The significance of this aerobic denitrification

    in nature is presently unknown.

    6.4.2. Influence of animals and plants

    Burrow-dwelling animals and rooted plants create a three-dimensional mo-

    saic of physico-chemical and biological microenvironments reaching deep

    into sediments. The surface area available for diVusive exchange and thearea of the oxicanoxic boundaries is considerably increased by the presence

    of these organisms (Aller, 1982) (see Chapter 4). Furthermore, the distribu-

    tion of reaction rates and solutes within the sediment varies in both time and

    space according to the patterns of activity of the plants and animals involved

    (Aller, 1994a; Christensen et al., 1994).

    Burrowing invertebrates such as polychaetes, crustaceans, and insect lar-

    vae are known to stimulate denitrification rates, measured per area of

    sediment surface, by a factor of 26 (Pelegri et al., 1994; Svensson, 1997).

    Irrigation of animal burrows results in advective transport of O2 and NO3

    from the overlying water to deeper anoxic sediment layers, where both

    nitrification and denitrification are stimulated (Kristensen et al., 1991b).some 30% of the NO3 produced by nitrification may be denitrified (Rysgaardet al., 1994), and higher percentages may be reached in less active sedi-

    ments with a thicker oxic zone (Seitzinger, 1988). At any given location,

    this percentage depends on the O2 concentration in the overlying water.

    Increasing O2 levels expand the O2 penetration depth, which decreases the

    NO3 diVusional loss of from the overlying water and favors the diVusion ofNO3 from nitrification into the denitrification zone (Figure 7.28).

    With high NO3 concentrations in the overlying water (eutrophic estuaries,lakes, and streams), total sediment denitrification rates are inversely propor-

    tional to O2 penetration depth. This is intuitive, as deep O2 penetration

    depths will also decrease the NO3 gradient into the sediment decreasingthe NO3 flux. This factor overrides the more modest increase in couplednitrification-denitrification rates as O2 penetration depth increases (Jensen

    et al., 1994). Oxygen regulates the genes that encode the enzymes required

    for denitrification, and gene expression is inhibited by oxygen concentra-

    tions of >1 mM (Zumft, 1997). Similarly, denitrification in the oceanic

  • 258 CANFIELD ET AL.However, they may not be stimulated equally, as Bartoli et al. (2000)

    reported that denitrification of water column NO3 (30 mM) was enhanced310 times more than coupled nitrification-denitrification in sediments

    Figure 7.28 (A), Total rates of nitrification and denitrification, and (B) denitrifi-cation as determined from 15N labeling in microcosm experiments on sedimentcollected from Lake Vilhelmsborg, Denmark. Indicated in (B) is whether denitrifica-tion is based on NO3 from the overlying water (Dw) or is coupled to nitrification-denitrification (Dn). Experiments were performed with diVerent O2 concentration inthe overlying water. The NO3 concentration in the overlying water was 30mM.Modified from Jensen et al. (1994).

  • 6.4.3. Temporal variations

    THE NITROGEN CYCLE 259Denitrification rates in aquatic sediments may vary on both seasonal and

    diel time scales, and the rates rarely correlate with temperature, which

    contrasts with other sediment respiratory processes such as sulfate reduction

    (see Chapter 9). The metabolic activity of denitrifiers in nature does correlate

    with temperature, and a Q10 response of around 2 is normal (from 0 to 35 8C;Focht and Verstraete, 1977). However, seasonal variations in denitrification

    rates are primarily regulated by the availability of organic carbon and NO3supplied by diVusion from the overlying water and by the intensity ofnitrification in the sediment. These parameters do not all vary in concert

    with temperature.

    Denitrification maxima are usually observed during the spring and occa-

    sionally, but to a lesser extent, during the fall in freshwater and marine

    sediments (Figure 7.29) (Rysgaard et al., 1995; Pattison et al., 1998). The

    distinct spring maximum is caused by a combination of increasing tempera-

    tures, ample supplies of NO3 from the overlying water, high rates of nitrifi-cation in the oxic sediment, and an increased supply of labile organic carbon

    from the microalgal spring bloom (benthic as well as pelagic). Low denitrifi-

    cation rates during summer are primarily related to lower NO3 and O2availability. Denitrification in sediments is frequently hampered as waterbioturbated by the polychaete Nereis succinea. Nitrogen-transforming

    processes are particularly enhanced when secreted mucus linings and asso-

    ciated organic particles along burrow walls provide labile organic carbon

    and nitrogen sources as well as reactive surfaces for microbial growth.

    Submerged and rooted macrophyte communities also stimulate area-

    specific denitrification rates in sediments. The degree of stimulation varies

    from near zero for the marine eelgrass (Zostera sp.) to more than sixfold in

    freshwater lobelia (Lobelia sp.) (Risgaard-Petersen and Jensen, 1997; Ottosen

    et al., 1999). Plants growing with roots in anoxic sediments have a well-

    developed lacunar system by which they establish a gas space continuum

    between leaves and the root tissue. Oxygen rapidly diVuses to the roots viathe lacunar conduits, where it supports the aerobic metabolism of the root

    cells. However, O2 may also diVuse across the epidermis of the roots and leakinto the surrounding rhizosphere. Consequently, a niche is formed for micro-

    bial communities performing coupled nitrification-denitrification. Species of

    macrophytes diVer in the amount of O2 released from their roots, rangingfrom almost nothing in Zostera sp. to large quantities in Lobelia sp. (Sand-

    Jensen et al., 1982). The denitrifier community in rhizospheres may be further

    supported by the concurrent release of easily degradable dissolved organic

    carbon from the roots.

  • 260 CANFIELD ET AL.column NO3 is scavenged by primary producers. Coupled nitrification-denitrification, on the other hand, may be O2 limited during warm summer

    months when high rates of microbial respiration create anoxic conditions

    near the sediment surface. Oxygen saturation is also reduced in warm water.

    Nitrification may be further suppressed due to competition for NH4 with a

    Figure 7.29 Denitrification in the sediment of the River Ouse over one seasonalcycle. Modified from Pattison et al. (1998).rapidly growing heterotrophic population. The reducing conditions prevail-

    ing in most organic-rich sediments during summer may also cause a partial

    shift from denitrification to NO3 ammonification. The latter process thrivesbest under strongly reducing conditions, when up to 80% of the reduced

    NO3 may pass via this pathway (Christensen et al., 2000).A small denitrification maximum in fall develops as a response to low

    temperature and reduced light conditions. Coupled nitrification-denitrifica-

    tion increases following the expansion of the oxic surface sediment, and the

    competition for NH4 by heterotrophs is reduced as overall microbial respi-ration rates diminish. The NO3 supply from the overlying water increases asprimary production drops and primary producers are unable to consume all

    the NO3 supplied. The low rates of denitrification prevailing during wintermonths are probably a direct temperature eVect because NO3 and O2usually are found in suYcient amounts.

    Diel variations in denitrification are observed only in shallow areas with

    dense populations of benthic microalgae. The high daytime primary produc-

    tion rates by microalgae increase O2 penetration into the sediment (Chapter

    13). The primary producers also actively consume available NO3 to supportbiomass production. Consequently, the supply of NO3 from the overlying

  • water is reduced, and NO3 originating from nitrification within the sedimentis increased (Rysgaard et al., 1994). Thus, during winter and spring, when

    most of the NO3 for denitrifiers is supplied from the overlying water, thedenitrification rate is lower during the day than at night. During summer, on

    the other hand, the rates are highest during the day when denitrification is

    primarily coupled to nitrification. However, competition for NH4 betweenbenthic microalgae and nitrifiers may hamper the latter process during the

    day when limited NH4 is available.

    7. ANAMMOX

    Table 7.8 Physiological and kinetic parameters of bacterial cultures duringaerobic (NH4!NO2 ) and anaerobic (anammox) (NH4!N2) NH4 oxidation in a

    THE NITROGEN CYCLE 261sequencing batch reactor

    Parameter Aerobic Anaerobic

    Free energy (kJ mol1) 272 357Biomass yield (mol mol1 C) 0.08 0.07Aerobic rate (nmol mg1 min1) 400 0Anaerobic rate (nmol mg1 min1) 2 60Growth rate (h1) 0.04 0.003

    Modified from Jetten et al. (2001).Nearly four decades ago Richards (1965) noted that NH4 did not accumu-late in O2 minimum zones of the ocean supporting denitrification, and he

    speculated that NH4 might be removed biologically by reaction with NO3 ,

    forming N2 gas, although no specific proof for this process was available.

    Somewhat later, thermodynamic considerations led Broda (1977) to the

    conclusion that the anaerobic oxidation of NH4 with NO2 (anammox) is

    as energetically favorable as oxic nitrification (Table 7.8), and he predicted

    the existence of chemolithoautotrophic organisms driving the following


    NH4 NO2 ! N2 2H2O 7:29Studies on the biological nature of the anammox process in wastewater

    bioreactors (van de Graaf et al., 1995; Jetten et al., 1999) have verified that

    N2 is indeed formed when one nitrogen atom from NH4 is paired with one

    from NO2 , as predicted by Broda (1977). Generation of N2 from NH4 and

    NO3 , on the other hand, requires more reducing power than available fromNH4 and is not in agreement with the stoichiometry of the anammox

  • 262 CANFIELD ET AL.process (Jetten et al., 2001). The source of NO2 for anammox in NO2 -poor

    environments is likely from NO3 reduction catalyzed by NO3 reductases

    and coupled to the oxidation of organic carbon.

    The only known chemolithoautotrophs involved in the anammox process

    belong to the order Planctomycetales (Strous et al., 1999a). Only a few

    species have yet been identified by enrichment and 16S rDNA sequencing;

    the first two are Brocadia anammoxidans, from a denitrifying pilot plant at

    Gist-Brocades, Delft, Holland, and Kuenenia stuttgartiensis, from a waste-

    water treatment plant in Stuttgart, Germany (Jetten, 2001). Organisms

    related to the wastewater species have been found in the water column of

    the Black Sea, where their distribution coincides with the zone of anaerobic

    ammonium oxidation (Kuypers et al., 2003).

    The metabolic pathway for anammox is not fully understood. Investiga-

    tions to date suggest that the electron acceptor NO2 is reduced to hydroxyl-amine (NH2OH), which reacts with the electron donor NH

    4 , leading to the

    production of N2 via the intermediate hydrazine (N2H4) (Jetten et al., 2001).

    At least three enzymes are believed to be involved in the process. The first

    enzyme reduces NO2 to NH2OH, the second catalyzes the condensation ofNH4 and NH2OH into N2H4, and the third enzyme catalyzes the oxidationof N2H4 to N2, thereby releasing the electrons required for NO

    2 reduction.

    Although the main product of the reaction is N2, about 20% of the con-

    sumed NO2 is recovered as NO3 . NO

    3 is probably formed as a byproduct

    when reducing equivalents are needed for CO2 reduction (van de Graaf et al.,

    1997). The anammox process is reversibly inhibited by O2 (Jetten et al.,

    2001), and no NH4 is oxidized at O2 concentrations >0.5% of air saturation(Figure 7.30) (Strous et al., 1997).

    The generation of N2 by anammox may be important in many aquatic

    sediments. Thamdrup and Dalsgaard (2002) demonstrated the anaerobic

    oxidation of NH4 with NO2 in marine sediments, presumably by the

    anammox reaction. They found that the relative importance of anaerobic

    NH4 oxidation for total N2 production in continental shelf sedimentsincreases significantly with water depth, from 2% in a eutrophic coastal

    bay (16 m) to 24% at 380 m and 67% at 695 m (Figure 7.31). The low relative

    significance of anammox in shallow sediments conceals that the absolute

    rates are two to three times higher than in sediments underlying deeper

    waters because the depth-dependent diVerence in denitrification is ordersof magnitude higher. The strong variation in denitrification is probably

    related to diVerences in the availability of electron donors, reflecting thedecreasing input and decreasing lability of sedimentary organic matter with

    increasing water depth (Liu and Kaplan, 1984). Thus, anammox has the

    potential to consume a significant fraction of NH4 produced in sedimentswhen only NO3 and NO

    2 are available, whereas denitrification is strongly

  • THE NITROGEN CYCLE 263dependent on the availability of reactive organic matter (Thamdrup and

    Dalsgaard, 2002). The process may also contribute substantially to N pro-

    Figure 7.30 The influence of oxygen on anammox activity. Cultures were incu-bated at the defined oxygen concentrations in a batch reactor. Activity could only bedetected when all oxygen was removed. Modified from Strous et al. (1997).2

    duction in NO3 -rich anoxic water columns (Dalsgaard et al., 2003). Hereanammox is dependent on heterotrophic NO3 reduction as a source of NH


    through ammonification.

    Anammox produces twice as much N2 per NO3 and NO

    2 molecule

    as denitrification. The anammox process can therefore explain nitrogen

    deficiencies in many anoxic waters and sediments, as well as the very eYcientconversion of NH4 to N2 in many shelf sediments, which has previouslybeen attributed solely to a tight coupling of nitrification and denitrification.

    The process must be included in future revisions of global nitrogen budgets,

    as it may fill significant gaps in the aquatic N2 production estimates.


    There are two stable isotopes of nitrogen, the major isotope, 14N (99.6337%),

    and the minor isotope, 15N (0.3663%). As usual, the isotopic composition of

    a sample is expressed relative to the minor isotope (Equation 7.30), and

    relative to a standard, taken as the (15N/14N) ratio of the atmosphere.

  • 264 CANFIELD ET AL.d15Nsam 100015N=14Nsam15N=14Nstd

    ! 1

    " #7:30

    As with many of the other biologically active elements undergoing redox

    transformations, isotope fractionation also accompanies the microbially

    mediated transformations of nitrogen compounds (Owens, 1987). Thus,

    assimilatory nitrate reduction, N fixation, denitrification, and nitrification

    impart distinct fractionations for potential use in reconstructing the relative

    intensity of these processes regionally and globally within aquatic ecosys-

    tems (e.g., Cline and Kaplan, 1975; Altabet and Francois, 1994; Sigman

    et al., 2000; Brandes and Devol, 2002), through the Holocene and Pleisto-

    cene (e.g., Francois et al., 1992; Farrell et al., 1995) and over geologic time

    (Beaumont and Robert, 1999). In the following sections we briefly overview

    the magnitudes of fractionation associated with the biological processing of


    Figure 7.31 Relative contribution of anammox to the total sediment N2 produc-tion (filled symbols), and the rate of sediment N2 production by anammox (opensymbols) as a function of water depth in Danish waters. The shallow station (16 m) isfrom Aarhus Bay, the intermediate station (380m) is from the southern Skagerrak,and the deep station (695 m) is from the central Skagerrak. Modified from Thamdrupand Dalsgaard (2002).

  • 8.1. Denitrification

    The isotope fractionation associated with denitrification has been widely stud-

    ied in pure cultures and natural populations of denitrifiers, and it has been

    inferred from the distribution of nitrate isotopic compositions in nature.

    From pure culture and natural population studies, relatively large fractiona-

    tions (eNO3-N2) of between 13% and 29% have been determined (Figure 7.32)(Delwiche and Steyn, 1970; Mariotti et al., 1981; Barford et al., 1999).

    Generally, these studies have been conducted under optimal growth conditions.

    Bryan et al. (1983) observed a negative relationship between the extent of

    fractionation and rates of denitrification (NO2 reduction to N2) for cells ofPseudomonas stutzeri under conditions in which NO2 was non-limiting andelectron donor availability was controlling denitrification. A similar trend

    was also observed for denitrification by cell-free extracts of P. stutzeri with

    abundant NO2 and limiting electron donor. When electron donors wereabundant and NO2 was limiting, both pure cultures and cell-free extracts

    THE NITROGEN CYCLE 265showed a positive correlation between rate and fractionation. In this case,

    NO2 availability controlled the rate. Bryan et al. (1983) concluded thatfractionation is not controlled by NO2 exchange across the cell membrane,as is partly the case for sulfate reduction (see Chapter 9). Rather, fraction-

    ation is apparently controlled by the availability of NO2 , as it controls thebuild-up of intermediates in the nitrite reduction process and the extent of

    back reactions. The fractionations observed during the denitrification of

    NO2 for whole cells ranged from about 5% to 25%. These fractionationsare similar in range to those observed during denitrification from NO3 (seeFigure 7.32).

    Figure 7.32 Isotope fractionations associated with N transformations by organ-isms. See text for details and references.

  • fractionations of between 18% and 36% (eNH4-NO3), while Mariotti

    in culture experiments with Trichodesmium sp. (strain IMS101), fractiona-

    tions (e ) of 1.3% to 3.6% were observed (Carpenter et al., 1997).

    266 CANFIELD ET AL.N2-biomass

    These fractionations are slightly higher than the fractionations of 0.5% to2% typically seen for Trichodesmium spp. in nature (e.g., Wada and Hattori,1976; Carpenter et al., 1997; Montoya et al., 2002). Although the fractiona-

    tions during N fixation are small, the isotopic consequences can be signifi-

    cant. The nitrogen source for N fixation is atmospheric N2, which, with a

    d15N of 0%, can have a significantly diVerent isotopic composition than thefixed nitrogen in aquatic systems. Nitrogen fixation, therefore, can produce

    isotopically distinct organic biomass. For example, the d15N of Trichodes-mium biomass in the surface oceans is typically 1% to 2%, with asso-ciated total particulate organic nitrogen (PON) in the range of 3% to2% (Carpenter et al., 1997). By contrast, surface water PON in areas ofet al. (1981) observed a large fractionation of 35% during nitrification

    by Nitrosomonas europaea. In a field study, the isotope and concentration

    profiles of NH4 in the water column of the Chesapeake Bay yielded model-derived fractionations during nitrification in the range of 13% to 17%(Horrigan et al., 1990). Taken together, nitrification produces substantially15N-depleted NO3 .

    8.3. Nitrogen fixation and assimilation

    Small but significant fractionations accompany the fixation of atmospheric

    N2 by microorganisms. Cyanobacteria have been the best studied, andThe isotopic composition of NO3 increases in oxygen minimum zones ofthe oceans (e.g., Cline and Kaplan, 1975; Brandes et al., 1998; Voss et al.,

    2001; Brandes and Devol, 2002). This isotope shift is indicative of denitrifi-

    cation, and modeling of the NO3 concentration and isotope profiles, typi-cally with a Rayleigh distillation model, produces fractionations in the range

    of 20% to 40%. These fractionations are similar to those found in pureculture studies (Figure 7.32). Denitrification in sediments, on the other hand,

    apparently produces only small net fractionations of around 3% (Brandesand Devol, 1997). This low fractionation is probably not due to low fractio-

    nations during the denitrification process, but rather to limiting NO3availability in rapidly metabolizing sediment systems.

    8.2. Nitrification

    Large fractionations also accompany the oxidation of NH4 with oxygen.In natural populations of soil nitrifiers, Delwiche and Steyn (1970) observed

  • the ocean devoid of N fixers is much more 15N enriched, with typical

    d15N values of 5% to 10% (Carpenter et al., 1997).Fractionations also accompany the uptake of nutrients by plankton

    and microorganisms and the subsequent processing of PON down the food

    chain. Phytoplankton fractionate (eNO3-biomass) from 3% to 10% as theytake up NO3 during growth (Montoya and McCarthy, 1995; Waser et al.,1998), with flagellates producing lower fractionations than diatoms

    (Montoya and McCarthy, 1995). When growing with NH4 , fractionationseems to depend on the NH4 concentration, and in experiments with diatoms,fractionations ranged from 8% at low NH4 concentrations of 5 to 20 mM to25% at concentrations between 50 and 100 mM (Figure 7.32) (Pennock et al.,

    THE NITROGEN CYCLE 2671996). High fractionations of up to 40% may also accompany microbial

    growth on NH4 , but these results arise from modeling studies in euxinicbasins and are not confirmed by direct determination (Velinsky et al.,

    1991). When grown on NO2 and urea, phytoplankton express only smallfractionations of around 0.5% to 1% (Figure 7.32).

    The extent to which high fractionations are expressed during nitrogen

    assimilation in nature depends on whether the nitrogen source is used to

    exhaustion during primary production. When the substrate is completely

    consumed, even large fractionations will not be expressed. The processing of

    PON through the food chain also influences the isotopic composition of the

    PON. Thus, the PON in animals is about 3.5% enriched in 15N, on average,

    compared to their food sources (e.g., DeNiro and Epstein, 1981; Montoya

    et al., 2002). The further up the food chain, the more 15N enriched the PON

    becomes. Isotope balance occurs as 15N-depleted NH4 is excreted by theorganism (Montoya et al., 2002). It is believed that the 15N enrichment of

    PON through the food chain is responsible for delivering 15N-enriched

    PON to the deep ocean, which, through subsequent oxidation, enriches

    deep-water NO3 in15N (d15N around 4.5% on average) relative to the

    primary-produced PON in surface waters (Montoya et al., 2002).

    The Nitrogen CycleIntroductionThe Global Nitrogen Cycle and Human PerturbationsBiological Nitrogen FixationThe Nitrogenase EnzymeAmmonium AssimilationThe Oxygen ProblemPhylogeny of Nitrogen-fixing OrganismsNitrogen Fixation in Aquatic Environments

    Microbial Ammonification and Nitrogen AssimilationAmmonificationDeamination and Ammonium IncorporationNitrogen Mobilization and ImmobilizationAnaerobic Nitrogen Mineralization and Ammonium Behavior in Sediments"New" Versus "Regenerated" Nitrogen in Pelagic Ecosystems

    NitrificationBiochemistry and Thermodynamics of NitrificationPhylogeny of Chemolithoautotrophic NitrifiersEnvironmental Factors Affecting Nitrification RatesSubstratespH and Surface GrowthTemperatureSalinityLight

    Dissimilatory Nitrate ReductionBiochemistry of DenitrificationNO3- reductaseNO2- reductaseNO reductaseN2O reductaseThe Entire Denitrification Process

    Biochemistry of No3- AmmonificationPhylogeny and Detection of DenitrifiersEnvironmental Factors Affecting DenitrificationDependence of NO3- and O2Influence of Animals and PlantsTemporal Variations

    AnammoxIsotope FractionationDenitrificationNitrificationNitrogen Fixation and Assimilation


View more >